Chironomids were collected during 24–27 February and on 10 November in 2018 in a stream connected to Mountain Lake, Virginia, USA. Mountain Lake is an oligotrophic lake that is located at an elevation of 1181 m above sea level near the summit of Salt Pond Mountain, and is the only natural lake of substantial size in the unglaciated part of the Southern Appalachian Mountain Range (Sharp 1933). Mountain Lake possesses an unusually high diversity of aquatic plants and invertebrates (reviewed by Parker 2003). Two species of Fontinalis have been recorded from the lake itself and the brooks in its vicinity, F. dalecarlica B. S. G. and F. novae-angliae Sull. (Studler 1982). F. dalecarlica is common in gently flowing water bodies, including in the brook that I sampled. Additional samples were collected during 4–9 April and 11–13 November in 2018 in a flowing-water stream, Rowans River, along Sparks Lane in Cades Cove, in the northwest part of the Great Smoky Mountains National Park (GRSM), Tennessee, USA, ca. 430 km away from Mountain Lake. Fontinalis novae-angliae with rigid stems and concave leaves occur from the streambed. Samples were collected in GRSM under research permit GRSM-2017-SCI-2389.

Insects were searched for underwater in the sampled streams and brooks and were collected together with the host plants occurring in their habitats. At Mountain Lake Biological Station, the clumps of mosses were detangled from detritus, and sediments were washed out of them. The larvae were placed in small plastic cases and observed with a microscope. The plastic chambers were constantly cooled with a refrigerant to keep their temperature in the range between 10–22 °C. Fourth-instar larvae were observed for 27 h in total, between 08:00 hrs and 18:00 hrs during 4–25 March 2018. Rearing and observations of chironomid larvae were performed at the National Museum of Natural History, Smithsonian Institution.

To compare and differentiate the chironomids of different stages and sexes occurring at the study sites, their partial COI (cytochrome c oxidase subunit I) gene sequences were determined. Total genomic DNA was extracted from 23 specimens at different stages, including adults (from a single adult leg or abdomen), larvae (two or three abdominal segments), and pupae (the whole abdomen) or pupal exuviae, using a NucleoSpin Tissue kit (Macherey-Nagel, Düren, Germany) and following the protocol provided by the manufacturer, with some modifications. The protocol was modified as follows: (i) tissue was digested for 48 h at 58 °C; (ii) after digestion with proteinase K, tissues were removed, washed in distilled water and used for morphological assessments; and (iii) the final elution volume was 30 μL. The primer pair used for the COI region consisted of primers 911 and 912 of Folmer et al. (1994), as used in some previous DNA barcoding and phylogenetic studies of Chironomidae (e.g., Guryev et al. 2001; Stur and Ekrem 2011; Cranston et al. 2012). Amplifications of the COI region were performed in a thermocycler, with an initial denaturation step of 94 °C for 4 min, followed by 40 cycles of 94 °C for 45 s, 55 °C for 45 s, and 72 °C for 1 min, and one cycle at 72 °C for 10 min. Amplification products were purified with ExoSAP-IT, according to the manufacturer’s instructions.

Direct sequencing of polymerase chain reaction (PCR) products was performed using the ABI Big Dye Terminator 3.1 cycle sequencing kit (Applied Biosystems, Lennik, Belgium) while following the manufacturer’s instructions and was carried out in an ABI 3130 Capillary Electrophoresis Genetic analyzer. Both DNA strands were sequenced. Sequences were deposited in the GenBank database (Table 1).

Sequence trace files were edited with 4Peaks v. 1.8 (by A. Griekspoor and Tom Groothuis, nucleobytes.com). Nucleotide sequences were aligned with Clustal W implemented in MEGA 7 (Kumar et al. 2016). To evaluate if the new species is phylogenetically exclusive among hypothetically closely related species (particularly the species in ‘devonica’ group, as discussed later) and to assess the intra- and interspecific genetic distances, the species delimitation plug-in in the software Geneious Prime 2019.2.3 (www.geneious.com) was used (Masters et al. 2011; Kearse et al. 2012).

The COI sequences of a rich record of Eukiefferiella species were found in GenBank with 1052 fragment sequences (accessed on October 10^th, 2019), although the sequence data identified at the species level were available only for five species (i.e., E. devonica (Edw.), E. ilkleyensis (Edw.), E. claripennis (Lundbeck), E. minor (Edw.), E. dittmari (Lehman)). In the dataset, 23 sequence data representing five species were included (Table 1), as well as five sequences obtained in this study. Phylogenetic trees were inferred by Bayesian inference (BI). Trees were rooted with two species of the genus Cardiocladius (i.e., C. capucinus (Zetterstedt), C. fuscus Kieff.). Evolutionary model was selected with MrModeltest v. 4.0b10 (Nylander 2004). The best fitting models were chosen with the Akaike Information Criterion (Akaike 1973). For the COI dataset, GTR + Gamma model was selected and used for the following Bayesian phylogenetic analysis. BI trees were constructed with MrBayes v.3.2.6 (Huelsenbeck and Ronquist 2001), using the plug-in of Geneious Prime, based on a cold chain and four heated chains with T = 0.2, running for 1,100,000 generations with a sample frequency of 200. The first 100,000 trees were discarded and the remaining trees were used to build a consensus tree, with estimated Bayesian posterior probabilities (PP). At the point of burn-in, the chains had all converged to a stable standard deviation of split frequencies lower than 0.01.

The species delimitation plug-in in the Geneious Prime 2019.2.3 (Kearse et al. 2012; Masters et al. 2011) was used (1) to measure the phylogenetic support of new species described herein, and (2) to evaluate the genetic differentiation among and within species, among other congeners which are hypothetically closely related to this species. For these purposes, Rosenberg’s P_AB (Rosenberg 2007) Rodrigo’s P (Randomly Distinct) (Rodrigo et al. 2008) were calculated. Rosenberg’s P_AB is a test for taxonomic distinctiveness of a clade based on the null hypothesis that monophyly is a chance outcome of random branching. Rodrigo’s P(Randomly Distinct) (‘Rodrigo’s P (RD)’) is the probability that a focal clade has the observed degree of distinctiveness (i.e., the ratio between the distance from a species-defining node to the tips of the tree, and the distance from that same node to its immediate ancestor) due to random coalescent processes (Rodrigo et al. 2008). Focal groups with values between 0.05 and 1 represent groups that have branching events that would be expected under the coalescent model in a Wright-Fishe population and a strict molecular clock. Additionally, six statistics which are useful to species delimitation were presented, along with Rosenberg’s P_AB and Rodrigo’s P (RD): the average pairwise tree distance, among members of the focal species/populations (‘D Intra’) and between the members of the focal species and members of the next closest species (‘D Inter’); as a measure of genetic differences between the focal species and its closest neighboring species, the ratio of D Intra to D Inter (‘Intra/Inter’); as the measures for evaluating diagnosability of each species/population, the mean probability of making a correct identification of a hypothetical sample of the focal species using placement on a tree under two different criteria, ‘P ID(Strict)’ (the sample must fall within the species clade) and ‘P ID(Liberal)’ (the sample is allowed to fall sister to or within a species clade); the mean distance between the most recent common ancestor of a species and its members (‘Av(MRCA)’). Accession number, voucher ID, and information on the localities for each specimen are shown in Table 1.

All specimens of adult abdomens, pupae, and larvae were digested with proteinase K, which made it relatively easy to examine the specimens morphologically. When necessary, the apical portion of the adult abdomen was macerated with warm (ca. 90 °C) 5% KOH and rinsed with distilled water. Each body part sample was mounted on a microscopic slide with Euparal.

The terminology of morphological features used herein followed Sæther (1971, 1977, 1980). The antennal ratio (AR), leg ratios (LR, BV, SV), wing ratio (L/WR) and hypopygium ratio (HR), and other morphological features were measured for adult male specimens following Sæther (1968), Schlee (1966) and Soponis (1977). Abbreviations that are used in this work:

AR antennal ratio: length of last flagellomere / length of remaining flagellomeres;

LR leg ratio: length of first tarsal segment/ length of tibia;

BV “Beinverhältnis”: length of femur, tibia plus first tarsal segment/ length of tarsal segments 2–5;

SV length of femur plus tibia/ length of tarsal segments 1–3;

L/WR wing length/ wing width ratio;

HR , hypopygium ratio: length of gonocoxite/length of gonostylus.

The type specimens are deposited in NMNH (National Museum of Natural History, Washington DC, USA). For each specimen, voucher ID is given as ‘YI-CR-##’.

Eukiefferiella represents a large and widespread genus of Orthocladiinae (Moubayed-Breil and Ashe 2013). This genus was erected by Thienemann (1926), and the independence of this genus from others, as well as the validity of its type species, has been debated in many studies (Edwards 1929; Thienemann 1936; Zavřel 1939; Brundin 1956; Lehmann 1972; Sublette and Wirth 1980). This genus is closely related to some orthocladiine genera in the Cardiocladius group (Sæther and Halvorsen 1981): Cardiocladius Kieff., Tokunagaia Sæth., Tvetenia Kieff., 1922 (Sæther and Halvorsen 1981), Nanocladius Kieff. (Sæther 1977; Lehmann 1972), and Hydrobaenus Fries. Keys for European species of this group are available for larvae (Thienemann 1936; Zavřel 1939; Chernovskii 1949; Bode 1983), pupae (Thienemann 1936; Lehmann 1972), adults (Brundin 1956; Lehmann 1972), and all these stages combined (Sæther and Halvorsen 1981).

The study of the taxonomy of Eukiefferiella is far from complete. Most existing records from the Nearctic are at the genus, or at most the species-group, level (Epler 2001). Ten species have been recorded from North America (Sæther 1969; Sublette 1970; Simpson and Bode 1980; Oliver et al. 1990; Bode et al. 1996; Patrick 1996; Sandlund and Aagaard 2004; Egan and Ferrington 2015): E. brevinervis (Malloch), E. brevicalcar (Kieff.), E. claripennis, E. coerulescens (Kieff.), E. cyanea Thienemann, E. devonica, E. gracei (Edw.), E. minor, E. unicalcar (Sæth.), and E. pseudomontana Goetghebuer. However, there is also evidently a large number of undescribed species (Cranston et al. 1983; Bode 1983; Merritt and Cummins 1996).

Eukiefferiella endobryonia sp. nov. was assigned to Eukiefferiella on the basis of diagnostic characters proposed by previous studies (Zavřel 1939; Lehmann 1972; Sæther and Halvorsen 1981; Cranston et al. 1983; Bode 1983; Sasa and Kikuchi 1995). Lehmann (1972) suggested that the possession of orally curved hooklets in a row located posteriorly on T III–V is a diagnostic character of the pupae of this genus, although he concluded that there is no single character by which this genus can be defined. Sæther and Halvorsen (1981) emended the definition of Eukiefferiella by comparing it to the other genera of the Cardiocladius group, and distinguished it based on the following features: lack of a male anal point; presence of a hind tibial comb; female gonapophysis VIII undivided; pupa lacking frontal setae, pearl rows, and a median anal seta; and larvae having a simple S I, seta interna, chaetae, and spinulae. Cranston et al. (1983) noted that Eukiefferiella is distinguished from Tvetenia based on a combination of the following larval characters: procercus less than 1.5 times as long as wide; abdominal setae shorter than 1/2 the length of an abdominal segment; and a simple S I.

Eukiefferiella endobryonia sp. nov. is unique in its lack of a precorneal seta, as most of its congeners have three precorneal setae present in a row or triangle (Lehmann 1972; Coffman et al. 1986); also, the configuration of pupal anal macrosetae (inner two macrosetae reduced, and outer one macroseta regular) is apparently unique for the species of Eukiefferiella. The affinity of this new species to others in this genus can thus be problematic. For example, E. endobryonia sp. nov. does not key out into any of the species groups proposed by previous authors (Brundin 1956; Lehmann 1972; Cranston et al. 1983; Bode 1983; Coffman et al. 1986). When compared to the species in Japan, this species may have affinity to chuzeoctava group defined in Sasa and Kikuchi (1995), although squama has only a few setae in this species (as opposed to squama fringed in the species of chuzeoctava group). Some morphological and behavioral traits of E. endobryonia sp. nov. show affinity to those of E. ilkleyensis in the E. devonica group, specifically: the larval mentum medially truncate with four lateral teeth; and the larva of E. ilkleyensis is reported to make a case at the shoot tips of the moss Eurhynchium riparoides (Hedw.) (Brennan and McLachlan 1980). E. ilkleyensis has been recorded from Palaearctic sites (Chernovskii 1949; Brennan and McLachlan 1979; Moller Pillot 2013; Erbaeva and Safronov 2016) but is not known from Nearctic sites. Hence, it is hypothesized that E. endobryonia sp. nov. can be included in the devonica group; importantly, however, this species morphologically contradicts with the distinguishing characters of the devonica group: the pupa of this species does not possess either median antepronotal setae or the row of hooks at the posterior edge of sternite VII (Lehmann 1972); deeply branched seta interna of the larva extending nearly to the base, although shallowly branched seta interna is said to be characteristic of the E. devonica group (Epler 2001); further, the larval thoracic setae of the new species are less than one-third as long as the segment from which they arise, as opposed to being more than half as long as the corresponding segment in E. ilkleyensis (Chernovskii 1949).

The molecular phylogeny (Fig. 3) clarified that E. endobryonia sp. nov. is genetically well-differentiated from E. ilkleyensis. This species is suggested to be relatively closely related to E. dittmari, which has not been assigned to any species group (Lehmann 1972) but may be incorporated in devonica group (Moller Pillot 2013). E. endobryonia sp. nov. is morphologically distinguished from E. dittmari based on the diagnostic character (i.e., the absence of precorneal setae and unequal length of pupal macrosetae in pupa). The immature stages of E. dittmari is not described, but the larvae live in mosses (Moller Pillot 2013). Based on the species delimitation analyses, the species clade of E. endobryonia sp. nov. was estimated to be distinctive and a significant level of genetic differentiation is detected between the geographic populations sampled (ML and GRSM). However, Rosenberg’s P_AB for each species used herein suggests that the null hypothesis that monophyly of each taxonomic group occurs due to the random coalescent process is not rejected. To clarify the phylogenetic status of the species and to trace the evolutionary history of biology of this group, especially the specialized relationship with mosses, it is necessary to elucidate the biology for the species of which immature stages are unknown and to accumulate morphological and genetic data of the related groups.

Eukiefferiella species are generally abundant in the riffles of rivers and streams (Madder et al. 1977; Brennan and McLachlan 1980; Egan et al. 2014), as well as in the littoral zones of lakes (Thienemann 1936). The species of this genus show a broad spectrum of life histories. Brennan and Mclachlan (1980) suggested that the larval feeding or tube-dwelling behaviors of species in this group may be affected by their habitats (i.e., riffle zones), as these are locations where the supplies of particles and sediments for tube construction are not always abundant.

Several species of Eukiefferiella are reported to be ectosymbionts of various freshwater invertebrates, including acting as parasites and commensals (Tokeshi 1995), although in many cases these ectosymbionts are not identified at the species level, or may not even have been described yet. An unidentified species in western North America makes silken tubes within the cases of the caddisfly Brachycentrus occidentalis Banks (Trichoptera: Brachycentridae), which may result in a higher mortality rate of the host (Gallepp 1974). E. aneyla Svensson and E. brulini Moubayed-Breil & Ashe are commensal on limpets of the genus Ancylus Müller (Gastropoda: Ancylidae) (Svensson 1986; Moubayed-Breil and Ashe 2013; Schiffels 2014; Fig. 4E). This genus also includes some species that are commensal on naucorid beetles (Roback 1977; Fig. 4H) or mayflies (Winterbourn 2004). Some species use rocks as substrates, including: E. verrallii (Edw.) and E. calvescens (Edw.), which construct loose tubes incorporating rocky particles (Brennan and McLachlan 1980); and E. clypeata Thienemann, which inhabits the exposed regions of rocks swept by water currents and builds solid sheets of silk (Brennan and McLachlan 1980) or cobweb-like tubes (Thienemann 1936; Chernovskii 1949) on the surfaces of stones, similarly to the filter net of hydrospsychid caddisflies. Cocoon-making is only known in one species, E. claripennis (Madder et al. 1977).

The larvae of Eukiefferiella are frequently found among aquatic mosses, including Fontinalis spp. (Zavřel 1939; Bode 1983; Suren 1993; Pinder 1995; Callisto et al. 2007). Interestingly, the fourth-instar larvae and pupae of E. ilkleyensis build tubes among the terminal leaves of the long-beaked water feathermoss Eurhynchium riparoides (Hedw.) (Hypnales: Brachytheciaceae) (Brennan and McLachlan 1980). E. ilkleyensis previously showed a preference for Fontinalis moss leaves as case materials (Brennan and McLachlan 1979). Feeding on living and decomposing leaves has been reported for some species (e.g., Callisto et al. 2007). For example, E. ilkleyensis is facultatively phytophagous on Ranunculus calcareus Butcher (Ranunculaceae) in its later larval instars (Berg 1995; Pinder 1995). The moss-feeding and leaf-rolling behaviors of E. endobryonia sp. nov. are apparently related to this species’ biological background. Some morphological traits of this species may be associated with their biology, including the secondary loss of the respiratory horns in the pupa. Respiratory horns represent a major morphological adaptation to oxygen-poor environments (Marziali et al. 2006; Marziali and Rossaro 2006), and tubicolous species often have variously shaped respiratory organs that lack a plastron plate (Cranston 1995). Therefore, the pupa of E. endobryonia sp. nov. lacks these additional respiratory adaptations and is dependent on cuticular respiration, although it should be noted that the tubicolous habit itself can be advantageous to respiration (Milne 1938; Walshe 1950; Williams et al. 1987; Kon and Hidaka 1983; Stief et al. 2005).

Building diverse, elaborate tubes is a characteristic behavior of the larvae of Chironomidae. Tube morphology is determined by construction behaviors, which are stereotypical for specific lineages (Chaloner and Wotton 1996; Charbonneau and Hare 1998). Tube construction behavior can therefore be seen as an example of these species’ “extended phenotype” (Dawkins 1982). Importantly, spinning in many chironomids is tightly linked with their feeding activity (Wallace and Merritt 1980; Dudgeon 1994; McLachlan and Ladle 2009). In addition, the case morphology of a species is determined by its habitat and the space, substrata, and particle sources available therein (McLachlan and Cantrell 1976; McLachlan 1977; McLachlan et al. 1978; Brennan and McLachlan 1979; Pinder 1986). However, the features of tubes that result from taxon-specific construction behaviors have often been studied with more emphasis on their food types and functional feeding group (Pinder 1986; Berg 1995; Ferrington et al. 2009). Virtually no previous study has ever given a comprehensive account of the diversity of chironomid tube morphology. Herein, I provide a provisional classification of chironomid tube structures based on their morphotypes (see below), with a summary of species’ biology and some examples. This enables us to give a brief overview of the biology of this group from the perspective of their construction behaviors to deepen our understanding of their ecology and evolution.

Chironomid tube structures can be categorized broadly based on their transportability and the substratum to which the tube is attached. First, tubes can be divided into those with fixed shelters (i.e., the larva cannot move around with the tube) and transportable cases (i.e., the larva freely moves while carrying the tube). Fixed shelters are much more common than transportable cases. Second, fixed shelters can be categorized into three groups (see below) based on the substratum to which the tube is attached. In fact, chironomids as a whole are able to colonize a broad spectrum of substrates. Third, these tube morphotypes can be further subdivided based on the materials out of which the tube is made, but only if the larva has a preference for certain types of particles (e.g., Brennan and McLachlan 1979; Xue and Ali 1997), and most tend to show low specificity for particular tube materials (Brennan and McLachlan 1979).

Tube structure is proximately influenced by the spinning mechanisms of larvae (Webb et al. 1981; Leuchs and Neumann 1990). The silk formation processes differ within the subfamily Chironominae (Webb et al. 1981; Tönjes 1989; Leuchs and Neumann 1990; Ólafsson 1992). Differences in silk properties exist between two subfamilies (i.e., Orthocladiinae and Chironominae) due to the influence of the striated ventromental plates, which are only present in the Chironominae (Churney 1940; Patrick 1966; Webb et al. 1981; Kullberg 1988). Orthocladiine larvae form separate threads of a soft, jelly-like substance, with which they spin loosely woven tubes that are normally attached to hard substrates (Webb et al. 1981). On the contrary, Chironominae manufacture a harder silk that comprises an amorphic liquid secretion while passing the ventromental plates and then becomes a sticky mass in water (Webb et al. 1981). They normally spin compact tubes, in which sediment particles are embedded in a silk matrix (Webb et al. 1981). In addition, some species can produce silk threads with different diameters (Walshe 1951; Leuchs and Neumann 1990), which may be achieved by shearing forces exerted by the ventromental plates or their tensile powers (Leuchs and Neumann 1990).

Chironomid tubes can most commonly be found bound to rocky materials in virtually all types of streams or lake ecosystems. These represent the first major morphotype of tubes: (1) rock material-bound tubes. Rock materials with various grain sizes are used to make tubes, ranging from very coarse (e.g., pebbles, cobbles) to fine (e.g., sand, silt, clay, mud). This type can therefore be subdivided by the grain size of particles used. Soil-dwelling chironomids often occur in lentic or lake environments, where they make soft, cryptic tubes with various lengths and forms, from short, cylindrical tubes to meandering, non-blindly ending tubes, in mud. The shapes of the silk-laden burrows or tubes of species dwelling in soft sediments are associated with their feeding strategies, as these are often deposit- and filter-feeding animals (McLachlan and Cantrell 1976; Leuchs and Neumann 1990; Stief et al. 2005). For example, a mud-dwelling species, Chironomus plumosus (Linn.) (Chironominae: Chironomini) (Fig. 5A), makes tubes with a variety of forms, including U- or J-shaped tubes, or horizontal tubes (e.g., McLachlan and Cantrell 1976; McLachlan 1977; LeSage and Harrison 1980; Hodkinson and Williams 1980).

It is noteworthy that three different spinning behaviors are known for species belonging to the genus Chironomus, which correspond to the different feeding strategies used by these species. The first method is found in the filter-feeding C. plumosus: the larva spins a funnel-shaped net across the lumen of the tube for filtering fine organic particles out of the water (Walshe 1947, 1951). The second method is another filter-feeding strategy, but is more common than the first type: the larva spins a thin layer of silk on the tube wall and then grazes on the food particles that attach to it inside the tube, together with the old net; the tube walls are thus renewed by the larva spinning new bundles of silk threads (Leuchs and Neumann 1990). The third method is deposit-feeding: the larva builds a cylindrical tube and reaches out of their tube to scrape up food particles from the surrounding sediments (Leuchs and Neumann 1990; Stief et al. 2005).

Some species of this type exhibit an aberrant, more elaborate construction behavior than the others. For example, larvae of Orthocladius (Euorthocladius) rivulorum Kieff. (Fig. 5B) make a suspended tube attached by one end to a riffle rock and feed on diatoms growing on the tube (Hershey et al. 1988; Taylor 1903; Peterson et al. 1993). They convert this tube into a pupal tube by binding the basal part with silk to form a long anchoring strand (Taylor 1903), which is similar to those structures made by limnocentropodid caddisflies (Wiggins 2005). The tubes of O. (E.) rivulorum are coated with diatoms, including Hannaea Patrick and Reimer (Fragilariaceae), and exert a considerable effect on the periphytic diatom flora (Pringle 1985; Hershey et al. 1988; Herren et al. 2017). Some Tanytarsini (Chironominae) make rigid tubes, such as Rheotanytarsus Thienemann & Baus and Lithotanytarsus Thienemann. Rheotanytarsus (Fig. 5C), which make tubes with arms on which the larvae spin silk strands attached to various substrates (Thienemann 1954; Scott 1967; McLachlan 1977; Kullberg 1988; Spies et al. 2009). Lithotanytarsus spp. and Rheotanytarsus reissi (Fig. 5D) spin serpentine tubes on the surfaces of travertine rocks, which become encrusted by carbonate mineral deposits, creating tufa over time (Thienemann 1934, 1936, 1954; Pinder 1995; Burmeister and Reiss 2003). Importantly, some ichnofossils have been interpreted as burrowing tubes from lentic environments (Chamberlain 1975; Rodríguez-Aranda and Calvo 1998; Sanz-Montero et al. 2013), with the oldest example of meandering tubes found in Triassic lake deposits (Voigt and Hoppe 2010). Calcareous columnar tufas resembling the biogenic constructions of Lithotanytarsus spp. have also been recorded in deposits from the Lower Cretaceous or later (Edwards 1936; Pentecost 2005; Brasier et al. 2011).

Figure 5. 

Summary of the biology of chironomids, with special focus on their tube structures. Some taxa without apparent tube structures (F, G) are included to give some accounts of their biology. Rock material-bound type (A–D): A three types of tubes, U-shaped (left), J-shaped (middle), and open-ended (right), built by the mud-dwelling species Chironomus plumosus, redrawn from McLachlan and Cantrell (1976) B larval (left) and pupal (right) tubes of Orthocladius (Euorthocladius) rivulorum, redrawn from Taylor (1903). C Rheotanytarsus rivulorum, redrawn from Walshe (1951) D Rheotanytarsus reissi, encrusted to form tufa, redrawn from Burmeister and Reiss (2003). Tubes on symbiotic animals (E–H): E Eukiefferiella brulini on Ancylus fluviatilis, redrawn from Moubayed-Breil and Ashe (2013) F Symbiocladius sp., an obligate parasite of the heptageniid mayfly Heptagenia lateralis, redrawn from Codreanu (1939) G Tonnoirocladius commensalis (Tonnoir), which is commensal to the larvae of the net-winged midge (Blephariceridae), redrawn from Tonnoir (1922) H silk tube of a Neotropical species of Eukiefferiella phoretically attached to Cryphocricos peruvianus De Carlo (Naucoridae), in which the open-ended tube is made between the meso- and meta-thoracic coxae of the host, redrawn from Roback (1977). Tubes on/within plants, algae, cyanobacteria (I–L): I silk tube made among algae by the marine species Clunio takahashii, shown in copula, redrawn from Hashimoto (1976) J a cone-shaped net constructed in the leaf mine of Endochironomus, redrawn from Walshe (1951) K a tunnel made by Cricotopus nostocicola in a spherical colony of cyanobacteria, Nostoc parmelioides, redrawn from Brock (1960) L a case formed by leaf-rolling on the shoots of Fontinalis mosses by Eukiefferiella endobryonia sp. nov., based on this study. Portable cases (M–P): M a solid tube composed of fine detritus constructed by Stempellinella minor, redrawn from Brundin (1948) N a tubular, firmly constructed case armored with diatoms and rhizopod shells constructed by Zavrelia pentatoma, redrawn from Lauterborn (1905) O a slightly curved, conifer needle case of Micropsectra pharetrophora, redrawn from Fittkau and Reiss (1998) P a laterally flattened purse case with concentric growth strips constructed by Lauterborniella agrayloides, redrawn from Lauterborn (1905).

Chironomids can also colonize a wide range of organisms, which include motile (animals) or sessile (e.g., plants, cyanobacteria) organisms. Information on the tubes formed by symbiotic chironomids is often limited, and their construction behaviors were not described in many previous studies. The second major group of tubes is: (2) tubes on symbiotic animals. Various chironomid lineages (Buchonominae, Podonominae, Chironominae, and Orthocladiinae) exhibit a wide range of symbiotic associations with aquatic animals, ranging from ectoparasitism to phoresis and commensalism, which can be either obligate or facultative associations (Steffan 1968; Tokeshi 1993, 1995; Schiffels 2014). Their hosts are mainly aquatic invertebrates, including snails (Prat et al. 2004; Fig. 5E), caddisflies (Ashe et al. 2015), mayflies (Claassen 1922; Tonnoir 1922; Codreanu 1939; Thienemann 1954; Jacobsen 1995; Fig. 5F), net-winged midges (Cranston 2007; Fig. 5G), naucorid beetles (Fig. 5H), sponges (da Silva Fernandes et al. 2019), and bryozoans (Moller Pillot 2009; Schiffels 2014). Host selectivity and the site of attachment to the host’s body are taxon-specific, and are likely determined by the behavioral patterns of the hosts and the ecological requirements of the chironomids, such as their feeding guilds (Tokeshi 1993). Some species that are tentatively categorized in this group herein presumably do not produce silk (e.g., Symbiocladius, Fig. 5F; Tonnoirocladius, Fig. 5G). The level of dependency on tube construction seems to differ greatly among taxa. For example, species of Symbiocladius (Orthocladiinae) (Fig. 5F) are obligate parasites of mayflies, and their larvae drill under the wing buds of the host and feeds on the host’s hemolymph. Symbiocladius spp. are not known to dwell in tubes, but may use other types of attachment mechanisms (Schiffels 2014). Some symbiotic chironomids, particularly facultative symbionts or those that are commensal to the host, tend to make tubes by weaving particles accumulated or flowing in the water on the body surface or shelters of hosts (Schiffels 2014). For example, Winterbourn (2004) gave an account of the tubes of an Eukiefferiella sp. in New Zealand, a commensal midge with a mayfly host, throughout its immature stages, and reported that it used silk thread to attach to the host; however, at the third and fourth larval instars, it built an open-ended silken tube above the mayfly’s gills that prevented them from beating, and the wall of this tube incorporated sand grains and fine organic particles.

Similar to the above, the third major category of tubes is: (3) tubes on/within plants, algae, cyanobacteria. Some chironomids in the Orthocladiinae and Chironominae often facultatively make tubes on the external surfaces of plants or algae (McGaha 1952; Berg 1995). For example, species of Telmatogon Schiner (Telmatogetoninae) inhabit intertidal zones and make their tubes within the green algae occurring on rocks (Cranston 2010). Clunio spp. (Orthocladiinae) are marine and make irregular silk tubes dotted with debris on seaweeds, and feed on encrusting diatoms and organic matter (Hashimoto 1976; Archer-Thomson and Cremona 2019). The female adult of Clunio takahashii (Fig. 5I) has a degenerated body and copulates with the male without fully extending her body out of her tube, and subsequently lays eggs in the tube (Hashimoto 1976). Many species in Chironominae and Orthocladiinae live inside plant tissues (Berg 1995), mainly as miners, but also rarely as gall-inducers (Jäger-Zürn et al. 2013). Some leaf-mining genera (Glyptotendipes Kieff., Endochironomus Kieff., and Polypedilum (Pentapedilum) Kieff.) are filter-feeders and spin a conical net inside the mine to filter out floating particles (Fig. 5J), similarly to Chironomus plumosus (Washe 1951). Species of Stenochironomus Kieff. feed either on living or sequestered plant leaves and tend to show weak preferences for host plants (Kato 2015). Cricotopus nostocicola Wirth (Orthocladiinae) (Fig. 5K) has a symbiotic relationship with colonial cyanobacteria in the genus Nostoc. The larva makes a tube by feeding on globular colonies of Nostoc. The feeding activity of the larva exposes the host to more sunlight, and consequently facilitates photosynthesis by the host (Brock 1960; Dodds and Marra 1989). For many miners (internal-feeders) of plants or algae, silk production has not been confirmed or described. This may be because they are more dependent on their substrates than other taxa. The tube of E. endobryonia sp. nov. (Fig. 5L), as described herein, exemplifies one of the few examples of a chironomid that make a tube exclusively using plant materials with silk. This study showed that the larva of this species dwells among the terminal leaves of moss plants, and draws their body out from the tube to consume all of the leaves occurring within a distance reachable by the larva; this species also forms a pupal case consisting solely of the apical five or six leaves of the moss, which are firmly attached together and internally lined by silk. It is interesting that a congener, E. ilkleyensis, has been reported to make a tube at the shoot terminus of mosses, yet without using silk for construction (Brennan and McLachlan 1979).

The fourth major group of tubes is: (4) portable cases. Among chironomids, transportable cases are far less common than fixed shelters. This is in sharp contrast to the myriad examples of portable cases seen among the Trichoptera (Wiggins 2005). Transportable cases have been found to be made by members of the Chironominae, mainly those in the tribe Tanytarsini, and less commonly in the tribe Chironomini (Walshe 1951; Reiss 1984; Ferrington 1995; Cranston 2010; Stur et al. 2005). Heterotanytarsus apicalis (Kieff.) is the only species so far known to the make portable cases among the species in the Orthocladiinae (Moog 1995; Moller Pillot 2013). Many genera of the subtribe Zavreliina in the tribe Tanytarsini, such as Stempellina Thienemann & Bause, Stempellinella Brundin (Fig. 5M), and Zavrelia Kieff. (Fig. 4N), live in cool mountain streams and construct heavy cases with hard shells composed of sand or silt (Lauterborn 1905; Hudson et al. 1990; Stur et al. 2005; Lenz 1924; Ekrem and Stur 2009; Namayandeh et al. 2016). Lauterborn (1905) described a cylindrical case made by Zavrelia pentatoma Kieffer & Bause (Fig. 5N), the surface of which was armored with diatoms and rhizopod shells; this species can also build cases exclusively out of rocks (Thienemann 1954). In the subtribe Tanytarsina (Tanytarsini), Calopsectra Kieff. and Micropsectra pharetrophora Fittkau & Reiss (Chironominae: Tanytarsini) (Fig. 5O) build ‘conifer-needle’ shaped case with spirally arranged sand grains, silt, and long pieces of detritus, especially diatom frustules (Fittkau and Reiss 1998). A few genera of Chironomini, such as Lauterborniella Thienemann & Bause and Zavreliella Kieff., also make motile cases with sand or silt particles (Thienemann 1954; Moller Pillot 2009). Lauterborniella agrayloides (Kieff.) (Fig. 5P) makes a soft purse- or spindle-shaped case that is slightly widened in the middle, strongly pressed together laterally, and rounded at both ends; it also contains characteristic concentric growth strips composed of numerous materials. This case resembles those of hydroptilid caddisflies (Lauterborn 1905). The construction behaviors and preferences for case materials of chironomids that make motile cases have largely not been investigated.

As seen above, many chironomid tubes are comparable to those of caddisflies. Not only are the forms of portable cases made by caddisflies and chironomids similar, but there are also notable shared characters of the pupal cases, known as ‘silken closures’ in caddisflies, between these groups (Wiggins 2005). Diverse forms of silken closures are known among many species of caddisflies in the suborder Integripalpia (Wiggins 2005). These silken closures are disc-shaped masses of silk that are used for closing an open end of the case, and are spun by the final-instar larva before entering pupation. Analogous structures, called ‘Verschlußdeckel (closure cap)’ or ‘Vorderdeckel (front cover)’ by Thienemann (1954), were described in some chironomids by previous studies (Lauterborn 1905; Lenz 1924). This represents a remarkable example of convergent evolution in portable case-makers in two distantly related insect clades (chironomids and caddisflies).

Tube morphotypes can be a tool for use in research in the taxonomy, ecology, and evolution of tube-building animals. The morphotypes of tubes should be assessed if they are useful for taxonomy. The tube, as a functional trait of a specific taxonomic group, may also be useful for examining different species’ ecological niches within a community (McGill et al. 2006). From an ecological perspective, tube functions have been and will continue to be an important subject at various scales, ranging from their adaptive significance for tube-dwelling individuals or species (e.g., Walde and Davies 1984; Williams et al. 1987) to their ecological impacts at the community or ecosystem scales (e.g., Ólafsson and Paterson 2004).

Herein, I made the simplistic assumption that the evolution of case-construction behavior in chironomids could be estimated based on the results of a previous molecular phylogenetic study (Cranston et al. 2012), although the phylogeny of this group has historically been contentious. Buchonominae was recovered as the most plesiomorphic subfamily (Murray and Ashe 1985; Cranston et al. 2012), although this contradicted some previous studies’ conclusions (e.g., Sæther 2000). Importantly, Buchonomyia spp. (Buchonominae) are ectoparasites of caddisflies, and supposedly do not manipulate silk (Ashe et al. 2015). The early diverging subfamilies identified, including Podonominae and Tanypodinae, are free-living (Moller Pillot 2013). Among extant lineages, Telmatogetoninae is likely to be one of the earliest-diverging clades of case-builders (Cranston et al. 2012). Tube diversity and construction behavior have evidently diversified in two more derived subfamilies, Chironominae and Orthocladiinae, which split during the mid-Jurassic. Interestingly, the spinning mechanism and silk properties are apparently different among the three clades of tube-making chironomids (Leuchs and Neumann 1990). The different ecological settings and evolutionary backgrounds of these groups have led to them developing a wide range of construction behaviors that have not yet been addressed within a geochronological framework. The emergence of two currently dominant and diverse clades of decomposers (i.e., caddisflies and chironomids) is estimated to have occurred in the Early Triassic (Cranston et al. 2012; Malm et al. 2013). Further, it has been suggested that the major clades of case-builders in both clades appeared contemporaneously during or after the Late Jurassic (Malm et al. 2013). The behavioral convergence of these tube-making insects may have been associated with major changes in freshwater ecosystems that occurred through geological time, such as increased allochthonous inputs of coarse plant materials and fine detritus into these systems (Lancaster and Downes 2013). It would be interesting to examine how these clades of ecosystem engineers have contributed to, or been affected by, the drastic changes in the trophic structures of stream and lake ecosystems that occurred in the Mesozoic Era (Lancaster and Downes 2013; Minter et al. 2016; Buatois et al. 2016; Savrda 2019).

The tube morphology of chironomids does indeed show significant diversity. It is likely that lability in silk production, particle and substrate selection, and construction behavior has made it possible for chironomids to use a broad array of ecological niches, as is the case for the Trichoptera (Weaver and Morse 1986; Wiggins and Mackay 1978; Mackay and Wiggins 1979; Wiggins 2005; Morse et al. 2019). Future investigations of the evolution and diversification of construction behaviors in chironomids in relation to changes in available habitat types will be important to improve our understanding of the evolution of aquatic ecosystems.