Introduction

Specialised larvae (glochidia) of the freshwater mussel order Unionida undergo a critical period during which they must parasitise suitable host fish to complete their life cycle (Kat, 1984). During reproduction, sperm aggregations (spermatozeugmata) are broadcast by unionid males into the water column and are carried by currents to nearby females, allowing for the uptake of sperm through their inhalant apertures (Ferguson et al., 2013). Fertilisation occurs internally, and fertilised embryos develop into mature glochidia within specialised brood chambers inside the gill demibranchs of the female (Bauer & Wächtler, 2001). Once mature, parasitic glochidia are eventually released into the water column and may survive only a few days before they must encounter and infect a host (Zimmerman & Neves, 2001; Hastie & Young, 2003; Melchior, 2017). Upon attachment to a suitable host, glochidia encyst and metamorphose into juveniles. These then drop from the fish and eventually settle, buried within a suitable substratum until they emerge at the surface several years later, developing into sexually mature adults (Bauer, 1988; Geist & Auerswald, 2007).

Strategies for transmitting parasitic glochidia to host fish are diverse in unionids (Haag, 2012). The type of release strategy employed by gravid female mussels is generally consistent with certain patterns of host fish use (i.e. specific host use with specialised lures), and unionids are often classified as either host fish infection generalists or specialists (Barnhart et al., 2008; Haag, 2012). Generalist glochidia may transform on a wide taxonomic range of fish species or feeding-guilds. Commonly, host generalists are found to simply broadcast their offspring into the water column, usually attached to mucus strands or webs that may serve to indiscriminately attach to or entangle potential host fish (Haag & Warren, 1998). Unionid species with host-specific glochidia infection strategies have a narrower immunological compatibility with host fish, being able to only metamorphose on a limited range of host species (Haag, 2012).

Specialist unionids have therefore evolved complex adaptations to increase the chance of attracting and attaching to an immunologically compatible host (Haag, 2012). For example, many specialists may use aggressive mimicry, a form of deception often used by parasites, to attract their target host species by exhibiting adaptations that imitate the host’s prey (Pasteur, 1982). In North America, in particular, examples of aggressive mimicry are commonly observed in gravid female unionids releasing parasitic glochidia. Specialist unionids may perform mimicry lure displays, using modified mantle margins that resemble host prey, such as larval and adult fish (Barnhart et al., 2008). Displays such as these have been observed to elicit attacks or feeding strikes, upon which the mussel expels its glochidia onto, or near the fish (Haag et al., 1995; Haag & Warren, 1998). Alternatively, glochidia release strategies may involve the production and release of conglutinates (mucus-aggregates containing glochidia) that mimic host prey items such as fish eggs and aquatic macroinvertebrates (Hartfield & Hartfield, 1996; Watters, 2002, 2008).

Studies also suggest that host species range may increase with glochidia size. Larger glochidia are often found on a wider range of hosts (e.g. Westralunio carteri Iredale, 1934, length ≥ 300 µm) and generally have shorter encystment times than smaller glochidia (e.g. Margaritifera margaritifera Linneaus, 1758, length ≤ 100 µm), where the length of encystment may range from weeks to months during which time glochidia also grow substantially (Bauer, 1994; Nezlin et al., 1994; Ziuganov et al., 1994; Klunzinger et al., 2013). Knowledge of the mechanisms of glochidia release and implications of glochidia morphometry are therefore critical for understanding species ecology and host–mussel interactions, and for developing conservation measures targeting threatened species of freshwater mussels.

Although widely documented around the globe, particularly in the Northern Hemisphere (Barnhart et al., 2008; Haag, 2012), unionid release mechanisms and morphometry of glochidia are poorly understood in New Zealand’s freshwater mussels (Bivalvia: Unionida: Hyriidae) (traditionally known as kākahi or kāeo). Two of the three species native to New Zealand, Echyridella menziesii and Echyridella aucklandica, co-occur in waterways in the northern half of the North Island (Marshall et al., 2014) where they have been ranked for conservation management as, At risk—declining (E. menziesii) and threatened—nationally vulnerable (E. aucklandica) (Grainger et al., 2018). Geriatric population-size structures are often recorded in New Zealand waters (e.g. Roper & Hickey, 1994), reflecting global trends of decline in mussel populations (Lopes-Lima et al., 2017) and inferring the existence of recruitment bottlenecks. Often, the conservation of declining species such as these is challenged by scant knowledge of their reproductive strategies and basic biology. In this paper, we report for the first time, glochidia morphometry and release strategies in E. aucklandica compared to E. menziesii. These data contribute to the basic understanding of the contrasting reproductive biology of two sympatric New Zealand freshwater mussel species, assisting future conservation interventions that facilitate population recovery, such as captive propagation and waterway restoration.

Methods

Mussel collection and maintenance

Gravid E. aucklandica and E. menziesii were hand-collected from populations in (i) the Ohautira Stream (37.762392, 174.98124), a short coastal stream in western Waikato, in November 2017 and January 2019, and (ii) the Mangapiko Stream (37.982022, 175.473541), a tributary of the Waikato River, in February 2018 and January 2019. Brooding female mussels were identified by in vivo examination and classification of gill demibranch pigmentation and volume (Table 1; Fig. 1). The procedure involves a non-destructive visual inspection using a nasal speculum to pry the mussel valves apart (ca. 0.5–1 cm). Gravid and fully charged (assumed to contain the entire brood) E. menziesii (Mangapiko n = 6, Mean (SE) length (L) = 60 ± 5.6 mm; Ohautira Stream n = 6, L = 57 ± 2.9 mm) and E. aucklandica (Mangapiko Stream n = 6, L = 77 ± 5.0 mm; Ohautira Stream n = 6, L = 84 ± 4.4 mm) were separated into buckets containing aerated 18°C water and sediment from the source location, and were then transported to The University of Waikato laboratory.

Table 1 Brood pouch characteristics used to determine gravidity stage by anatomical examination (gill colour and volume) of sampled Echyridella menziesii (modified from Melchior, 2017) and Echyridella aucklandica
Full size table
Fig. 1
figure 1

Marsupia of female Echyridella menziesii (1A outer valve, 1B gravidity stage 3, brooding unviable glochidia 1C gravidity stage 4, brooding viable glochidia; 1D gravidity stage 5, spent) and Echyridella aucklandica (2A outer valve, 2B gravidity stage 3, brooding unviable glochidia 2C gravidity stage 4, brooding viable glochidia; 2D gravidity stage 5, spent) at various stages of gravidity, complementing Table 1

Full size image

Mussels were held within 5-l aerated aquaria containing dechlorinated tap water and three cm depth of silica sand (one mussel per aquarium). Ten percent of the water in each aquarium was changed every other day to minimise build-up of ammonia and other waste products. Mussels were fed a mixture of Reed Mariculture Nanno 3600 and Mariculture Shellfish diet diluted with 1-l dechlorinated tap water to provide ca. 4700 cells/ml/mussel/day (Ganser et al., 2015). Controlled temperatures (18°C) and a light:dark cycle (16:8 h) matched ambient conditions, allowing adult females to release their broods naturally.

Glochidia release

Adult E. menziesii and E. aucklandica were observed daily for release of glochidia. If glochidia were found to be released, these were retrieved using a 5-ml pipette and placed on a petri dish or watch glass under the Olympus SZ-6045 dissecting microscope to determine whether they were released as individual glochidia, embedded in mucus strands, or as conglutinates. Samples of glochidia were analysed for maturity, characterised by (i) the presence of hooks on opposing valves, (ii) translucent valves, free of their vitelline membrane, and iii) rapid opening and closing of the glochidia valves. All glochidia collected from each mussel were preserved in 70% ethanol for later analysis. If the sample contained conglutinates, distinctions were made between puerile or non-functional (containing only premature glochidia) and functional (structures containing mature glochidia) conglutinates (Barnhart et al., 2008). Maturity of glochidia was assessed as above. All conglutinates released by each mussel were preserved in 10% formalin.

Fecundity was defined as the total number of glochidia brooded by a female during a single brooding event, with the assumption that mussels produced one clutch per year. Estimations of fecundity were made using the collected glochidia or conglutinate content from three females per species per stream population. If at the completion of the trial, mussels had not released their entire gill contents, these were then flushed using dechlorinated tap water to remove remaining glochidia material. Echyridella menziesii glochidia were diluted to a 100-ml homogenised dechlorinated tap water solution. The number of glochidia was then counted in five aliquots of 1-ml under the Olympus SZ-6045 dissecting microscope. Fecundity was estimated by multiplying the mean number of glochidia in the five sub-samples by dilution volume. Echyridella aucklandica fecundity estimations were made by counting glochidia attached to a sub-sample of five conglutinates and multiplying the mean number of glochidia by total released conglutinates (excluding conglutinate fragments).

Glochidia valve morphometry

A sub-sample of preserved glochidia was photographed and measured using the Leica DM RD equipped with the Olympus DP70 Digital Camera system and Olympus image analysis software. Length (widest part of the shell between anterior and posterior edges, parallel to the hinge), height (widest part of the shell perpendicular to the hinge) and hinge length were measured to the nearest 1 μm (see Fig. 2) to compare morphometric differences among individuals, populations and species. For E. menziesii, 10 glochidia from 3 adult mussels per stream population were measured for a total sample of 60 glochidia. For E. aucklandica, 3 glochidia from 3 conglutinates of 3 mussels were measured for a total of 27 glochidia per stream population (total n = 54). Height, length, size (average of height and length, Barnhart et al., 2008) and height/length ratio (shape) were compared among populations and species using t-tests for independent groups (STATISTICA version 13; Stat Soft Inc. Oklahoma, U.S.A.).

Fig. 2
figure 2

Morphometric measurements of mature A Echyridella menziesii and B Echyridella aucklandica glochidium valves: A–B, length; C–D, hinge length; E–F, height (modified from Klunzinger et al., 2013). Anatomical orientation descriptions are labelled according to Hoggarth (1987)

Full size image

Results

Echyridella menziesii marsupia were found to be positioned in the inner one-third of the inner demibranchs. The marsupia increased in volume with developmental stage and changed in pigmentation from transparent yellow with immature glochidia (Stage 3) to dark orange, containing mature glochidia before glochidial parturition (Stage 4, Table 1, Fig. 1). Echyridella aucklandica had marsupia located at the posterior end of the inner demibranchs. These occupied a greater surface area (two-thirds) of the gills than those of E. menziesii, and the pigmentation of the demibranchs in E. aucklandica ranged from orange and purple (Stage 3) to fully purple (Stage 4) for gravid females (Table 1, Fig. 1).

Glochidia release strategies

Within laboratory aquaria, all female E. menziesii from Mangapiko and Ohautira Stream populations exhibited dual host infection methods by releasing translucent glochidia both individually and bound to mucus threads (up to 4–5 cm in length) attached temporarily to the exhalant aperture of the adult (Fig. 3). Rhythmic contractions of the adult female’s exhalant aperture were observed to cause the mucus threads to drift and suspend glochidia in the water column, as shown in the video (Online Resource 1). Glochidia strands often became entwined with one another, leading to mucus masses containing glochidia that were observed lying on the sediment next to adult female E. menziesii.

Fig. 3
figure 3

Adult Echyridella menziesii releasing glochidia bound to mucus strands attached to the exhalant siphon (white arrow) (A), and photo-microscopy of mature and viable hooked E. menziesii glochidia (B)

Full size image

All E. aucklandica from both the Mangapiko Stream and Ohautira Stream populations released glochidia attached to conglutinates (Fig. 4; see video, Online Resource 2). Individual conglutinates were expelled through the exhalant aperture, and mussels were sometimes observed to discharge conglutinates that were attached temporarily to an adhesive mucus strand protruding from between the extended tissues of the mantle margin (as shown in the video: Online Resource 3). Detached conglutinates (intact and fragmented) were then observed floating with wavelike motions in water currents created by aeration or tumbling on the bottom of the tank.

Fig. 4
figure 4

Two conglutinates released by adult Echyridella aucklandica (A). Photomicrograph of E. aucklandica conglutinate (L: 6 mm) containing mature and viable glochidia (B). Compound photomicrograph of E. aucklandica glochidia attached to the outer layer of the conglutinate (C). Puerile conglutinate containing unviable (closed) glochidia (D)

Full size image

Echyridella aucklandica produced two types of conglutinates: functional conglutinates, which contained mature glochidia attached externally to the conglutinate material (Fig. 4B, C); and puerile (or non-functional) conglutinates (Fig. 4D). Very few puerile conglutinates were released by adults from both populations (n  ≤ 5–10 per mussel), commonly at the end of the release period. Puerile conglutinates differed from functional conglutinates in shape, size and colour. These structures were entirely brown material containing immature glochidia (encased in vitelline membranes) embedded within the conglutinate material. Lengths for puerile conglutinates for mussels from both Mangapiko and Ohautira Stream populations ranged from 5 to 10 mm.

Functional conglutinates varied in morphology and size (L: 4–9 mm; W: 2–3 mm) from both Mangapiko and Ohautira Stream populations; however, all were dorso-ventrally flattened with a vermiform appearance (Fig. 4A, B). All functional conglutinates were composed of solid, spongy mucus and shared similarities in colour, with a white membrane and the dorsal surface partly covered with tan-coloured material to which translucent mature glochidia were attached by their hinge or outer valves (Fig. 4B, C). Functional conglutinates were spontaneously discharged by individuals one-at-a-time at a rate of 12–21 per day over a period of 1–2 weeks. At times, releases were also observed as single conglutinates, occurring in response to stimulation by touch of the pipette (reflexive release; Barnhart et al., 2008). The number of functional conglutinates released ranged from 49 to 239 per mussel (\(\overline{x}\) = 136.5 ± 59.4 SE) for both populations (n = 12) with 141 ± 50.6 for Mangapiko adults (n = 6) and 131 ± 36.8 for Ohautira adults (n = 6). Fragments of conglutinates were not quantified.

The number of glochidia per conglutinate (n = 15) ranged from 61 to 280 (\(\overline{x}\) = 173.7 ± 19). Estimated fecundity was found to be greater in E. menziesii (n = 6; L 59 ± 3.6), with some females producing broods that contained nearly twice the number of glochidia (\(\overline{x}\) = 44,016; range 28,840–72,000) than E. aucklandica (n = 6; L 85 ± 1.8), which only released an estimated mean of 17,840 glochidia (range 1737–34570), although this difference was not statistically significant (t(11) = 2.5, p = 0.3). No significant differences in fecundity were found between Ohautira and Mangapiko Stream populations in E. menziesii (t(5) = 0.4, p = 0.7) or E. aucklandica (t(5) = 0.5, p = 0.6).

Glochidia description

Released glochidia from both E. menziesii and E. aucklandica were sub-triangular in shape with a straight hinge and hooks proximal to the apex of each valve. Valves of E. menziesii glochidia compared to E. aucklandica were distinctly more inequilateral (i.e. the ventral hooked edge of each valve was off-centre and displaced posteriorly with the anterior edge being longer and having a more prominent curve) (Fig. 2). Glochidia shape (shell height:length ratio) differed significantly between the species (t(112) = 14.2, p < 0.01), with E. menziesii glochidia having a lower height-to-length ratio (0.83 ± 0.02), resulting in more elongated valve shapes compared to those of E. aucklandica (0.92 ± 0.01) which had an almost equal height-to-length ratio. Further significant differences between species were found in glochidia shell size (mean shell height and length, Barnhart et al., 2008), with E. menziesii (277 ± 0.7 µm) nearly 3 times the size of E. aucklandica (99 ± 0.3 µm SE) (t(112) = 220.9, p < 0.01). When comparing glochidia valve sizes within species between Mangapiko and Ohautira Stream populations, no significant differences were found (E. menziesii: t(59) = 1.3, p = 0.19; E. aucklandica: t(53) = 0.44, p = 0.66). However, E. aucklandica glochidia were found to have significant differences in shape between stream populations, with Mangapiko Stream populations having a greater length-to-height ratio than glochidia from Ohautira Stream populations (t(53) = 3.51, p  ≤ 0.01). Echyridella menziesii valve shapes were found not to be significantly different (t(59) = 1.3, p ≤0.20) between stream populations (Table 2). No larval threads were observed in any of the studied E. menziesii or E. aucklandica glochidia samples.

Table 2 Mean (± standard error; ranges in parentheses) length (L), height (H) and hinge length of glochidia and size measured in µm
Full size table

Discussion

Host infection strategies

Host infection strategies described in this study for both Echyridella aucklandica and Echyridella menziesii represent a contrasting set of behavioural and morphometric adaptations for the two species to facilitate transmission of glochidia to hosts. To our knowledge, the Australasian hyriid, E. aucklandica is the first non-North American mussel species, reported to use a host attraction strategy through the release of functional conglutinates that resemble vermiform macroinvertebrate prey items of fish, such as some Diptera larvae, Hirudinea or Turbellaria. In North America, host-attracting, functional conglutinates are a common feature in unionids (Haag, 2012), whereas in Europe, the conglutinate production reported to date appears to be non-functional (consisting of unviable glochidia or eggs and not serving as a host attractant) and induced by stress (Aldridge & McIvor, 2003; Lopes-Lima et al., 2017). In contrast, the functional conglutinates released by E. aucklandica were embedded with mature and viable glochidia, suggesting that the structures are adapted to attract and increase infection rates on fish (Haag, 2012). Although structurally variable, all E. aucklandica conglutinates were dorso-ventrally flattened and composed of solid, spongy mucus bodies to which glochidia were attached. The North American Creeper mussel Strophitus undulatus (Say, 1817) produces a similar type of conglutinate, described as a translucent, milky, rod-shaped (3–7 mm) conglutinate that contains 1–15 glochidia, probably mimicking maggots and other insect larvae (Haag, 2012; Watters, 2002, 2008). Watters (2008) categorised these structures by morphology and composition into the meso-conglutinates (mature larvae attached to solid mucus conglutinates). The description of the meso-conglutinate produced by S. undulatus parallels with the description of E. aucklandica conglutinates within our study. However, in contrast to E. aucklandica which was found to produce miniature glochidia, S. undulatus is known to release the largest glochidia known (> 200–500 µm) and unlike other conglutinate producers, S. undulatus are host generalists, using at least 15 unrelated hosts for the transformation of their glochidia (Watters, 2002). Echyridella aucklandica conglutinates are also similar in shape to the leech like (2–5 mm) conglutinates of the Dromedary pearly mussel (Dromus dromas Lea, 1834) (Jones et al., 2004). However, unlike the E. aucklandica conglutinates, which did not display colour variability and only contained viable glochidia, D. dromas conglutinates range in colour from red to white and contain immature eggs that hold together the centre of the conglutinate with mature glochidia on the outer margin (Jones et al., 2004), and are thus grouped into composite conglutinates (Watters, 2008). Glochidia released by D. dromas have been successfully transformed on ten benthic-feeding fish species in the laboratory, including nine species of darter (Percidae) and one sculpin (Cottidae).

Haag et al. (2012) divide North American unionid conglutinate releasers into several sub-categories: (i) pelagic, (ii) mucoid and (iii) demersal. Pelagic conglutinates are forcefully expelled into the water column and immediately enter the drift, targeting drift-feeding fish such as minnows (Fusconaia sp.; Patterson et al., 2018). Most pelagic conglutinates seem to be generalised mimics that do not resemble specific organisms but rely on their motion in stream drift to attract fishes (Haag, Center for Mollusk Conservation, Kentucky Department of Fish and Wildlife Resources, pers. comm.). Mucoid conglutinates are associated with benthic habitat, and do not resemble specific prey groups, lacking a clear structure. These conglutinates are often found in species that target generalist feeders as a host (e.g. Cyclonaias pustulosa (Lea, 1831) targeting catfish, Patterson et al., 2018). Demersal conglutinates, on the other hand, are structured packages of larvae that can be very elaborate. These packages resemble a variety of food items, including leeches, worms and aquatic insects such as Ptychobranchus blackflies. Conglutinates either quickly settle to the bottom, stay adhered to the parent mussel, or in the case of Cyprogenia sp. stay attached to the mussel for a period of time before settling to the substrate. The demersal conglutinate release strategy is strongly oriented towards parasitism of benthic host fish, targeting small invertivores (Patterson et al., 2018).

Our laboratory observations of E. aucklandica conglutinate release suggest the potential for both pelagic and demersal release strategies. When first released, conglutinates were seen floating with a rippling motion in the aquaria, which may be a useful strategy to attract pelagic or drift-feeding fish. After a short time, however, conglutinates settled to the bottom of the tank in the absence of flow, which would place conglutinates in prime locations to be preyed upon by benthic feeders in the wild (Patterson et al., 2018). The use of conglutinates as a host attraction strategy for E. aucklandica, in comparison to a passive attachment strategy for E. menziesii, suggests that E. aucklandica glochidia may require different host fish species to E. menziesii and that E. aucklandica’s glochidia may parasitise internal structures due to possible consumption by attracted hosts. Furthermore, E. aucklandica may be a potential host fish specialist as other studies of species using similar release strategies suggest that these evolved to suit the feeding habits of a host species or a suite of fish feeding-guilds adapted to attract and facilitate glochidia transfer to a specific feeding guild (Barnhart et al., 2008; Haag 2012; Patterson et al. 2018). Field research is currently underway to identify host fish infection by E. aucklandica and E. menziesii, with initial findings identifying pelagic species of Retropinnidae and Galaxiidae as host fish for E. aucklandica but not benthic Anguillidae or Eleotridae (Melchior & Hanrahan, The University of Waikato, unpublished data, February 2019).

It is important to note that the observations of glochidia release behaviour in this study have been of mussels held in a laboratory setting, and release behaviours may vary in the wild. Previous studies found that some unionids (e.g. Quadrulini sp.) produce puerile conglutinate-like structures, often composed of eggs or developing embryos (Barnhart et al., 2008) that may be released as a stress response under hypoxia (e.g. Unio pictorum Linnaeus, 1758) and Unio tumidus (Philipsson, 1788) (Aldridge & McIvor, 2003). In the present study, aerators were used in all aquaria to reduce stress to adult mussels and only a small number (< 5 per mussel) of puerile conglutinates were released. Ideally, future studies would include field observations of conglutinate release to fully elucidate host attraction and infection strategies for this New Zealand unionid species.

Unlike E. aucklandica, which uses mimicry in the form of conglutinates as a strategy presumed to attract host fish, E. menziesii uses a passive form of infection by broadcasting both individual glochidia, and glochida-laden threads of mucus that probably serve to suspend larvae into the water column and entangle passing hosts. Passive entanglement is most often found in generalist mussel species (including Australasian hyriids Westralunio carteri (Haag, 2012; Klunzinger et al., 2013) and Hyridella drapeta (Iredale, 1934) (Jupiter & Byrne, 1997), as a variety of fish may be infected indiscriminately through contact with the released mucus webs (Haag & Warren, 2003). Echyridella menziesii is now well known to parasitise a range of native New Zealand fish, and can thus be categorised as a host-generalist based on existing information. In the wild, glochidia have been found on fish species such as kōaro (Galaxias brevipinnis Günther, 1866)  and small benthic-feeding fish such as the giant bully (Gobiomorphus gobioides Valenciennes, 1837) (Percival, 1931; Hine, 1978). More recently, successful laboratory glochidia attachment and transformation was demonstrated on rainbow trout (Oncorhynchus mykiss Walbaum, 1792) (Clearwater et al., 2014), common bully (G. cotidianus McDowall, 1975), native banded kōkopu (Galaxias fasciatus Gray, 1842), shortfin eel (Anguilla australis Richardson, 1841), longfin eel (A. dieffenbachii Gray, 1842), and Canterbury galaxias (G. vulgaris Stokell, 1949), with the highest transformation rates on common bully (91%) followed by banded kōkopu (69%) (Brown et al., 2017). However, host infection and transformation success is currently unknown for E. aucklandica.

In this study, estimated fecundity varied within and among species, which is a consistent finding in unionids generally (Haag & Staton, 2003). In North American mussels, the number of glochidia produced annually ranges from < 2000 up to 10 million (Haag, 2013). Body size has been suggested to be a strong predictor in fecundity of freshwater mussels within and among species, with large individuals generally found to have a higher reproductive output (Haag, 2013). This is inconsistent with the data collected within this study, although sample size was small. Based on the material collected, estimated clutch size per adult was considerably lower in the larger-sized E. aucklandica in comparison to E. menziesii. Furthermore, both species produced a considerably lower number of offspring than some North American mussels (Haag, 2013) and the hyriid H. drapeta (Byrne, 1998). The low fecundity estimated in both species may reflect complex relationships with host fish due to the efficiency of host infection through the production of conglutinates in E. aucklandica and the wide range of hosts available for infection for the host generalist, E. menziesii. Haag (2013) indicates that the release of conglutinates and mucus webs are generally associated with low length-standardised fecundity. The production of conglutinates and release of puerile conglutinates (containing immature glochidia) may significantly increase energetic costs and therefore lower fecundity, whereas the production of mucus strings associated with host generalists increases the chance of glochidia infection, allowing for low fecundity.

Glochidia morphometry

Glochidia of both Echyridella species were morphometrically distinct from each other in size and shape. Echyridella menziesii glochidia were larger than E. aucklandica and distinctly larger than the size reported for most Australasian Hyriidae, with the exception of Westralunio carteri (309 µm) (Klunzinger et al., 2013) and Hyridella drapeta (320 µm) (Atkins, 1979). The ability to identify species in the field using morphometrics is important, particularly in riverine systems where species co-occur, as is the case for E. menziesii and E. aucklandica. Kennedy & Haag (2005) used morphometrics to identify 72–79% of total glochidia from the Sipsey River system that is inhabited by 21 unionid species. As with our research, the study found low within-population variability, with the exception of one species, supporting the general use of morphometry to differentiate between unionid species. Though within-population variability for glochidia morphometry was low for our study—variability between populations seems to occur throughout New Zealand with glochidia dimensions of E. menziesii in this study being distinctly smaller than in previous studies by Percival, (1931) and McMichael & Hiscock (1958) who measured glochidia lengths of 360 µm (Lake Sarah, West Canterbury, South Island) and 310 µm (Waikato River, Hamilton, North Island), respectively.

Small glochidia, such as those produced by E. aucklandica, have been reported for other unionid species (e.g. the pearl mussel Margaritifera margaritifera—size < 100 µm). These were observed to attach to the gills of host fish, whereas larger glochidia are found primarily on the fins of their hosts (Bauer, 1994). Attachment to skin and fins is commonly found in triangular-hooked glochidia of Unionidae and Hyriidae (Bauer & Wächtler, 2001). The combinations of larger size and a low height:length ratio, as found in E. menziesii, are traits that improve leverage and gripping force (Hoggarth & Gaunt, 1988), allowing the glochidia to easily attach externally to the fish (Bauer, 1994). Furthermore, larger glochidia have been reported to have shorter host retention times than smaller glochidia that may be retained for a longer period, enabling them to not only develop internal organs but also grow in size while on their host (Bauer, 1988; Nezlin et al., 1994).

Host specificity is associated with selective encounters of host fish taxa or due to the dominance of a particular host species in certain habitats. The relationship between unionids and host fish is easily disrupted, particularly in New Zealand, where a significant proportion of the potential host fish pool is diadromous, adding to the risk of recruitment disruption for E. aucklandica in particular. Contrasting glochidia release and attachment strategies by both species suggest that sympatric populations of E. menziesii and E. aucklandica may co-exist through partitioning of host resources. Furthermore, differences in release strategies and morphology in both glochidia and adults also suggest that E. menziesii and E. aucklandica may not be as closely related as previously reported and should be further investigated using integrative approaches that combine molecular, morphological and ecological data. As well as providing the basis for recognising and distinguishing glochidia of these two Echyridella species in field studies of host fish infection, our novel finding of probable host fish attraction through mimicry of conglutinates in E. aucklandica highlights the importance of understanding the fundamental biology of host–mussel relationships and the need to develop species-specific methods for captive propagation to support restoration of declining populations.