Introduction

Unionoid populations today suffer from many, commonly anthropogenic factors and are in serious decline worldwide (e.g., Ferreira-Rodríguez et al., 2019), thus the possibility of obtaining any biological knowledge of their rarely observed juvenile life stage that is also believed to suffer high mortality is essential for both establishing and optimizing conservation strategies as well as addressing functional or evolutionary questions. Common freshwater mussels classified in the order Unionida (unionoids, naiads) form a monophyletic and biologically coherent taxonomic unit. They live today in lakes and rivers worldwide (Bogan & Roe, 2008; Lopes-Lima et al., 2018; Graf & Cummings, 2021). They are also currently regarded as one of the most threatened freshwater taxa in the world (e.g., Haag & Williams, 2014), with estimates of up to 985 living species worldwide (Bogan & Roe, 2008; Graf, 2013; Graf & Cummings, 2007, 2021, 2023), with their diversity decreasing rapidly on a global scale (Seddon et al., 2011; Haag & Williams, 2014; Pereira et al., 2014; Lopes-Lima et al., 2018; Van Tu et al., 2018; Ferreira-Rodríguez et al., 2019; Haag, 2019; Böhm et al., 2020). These bivalves are characterized by a unique life cycle: their larvae must pass through the parasitic stage on a fish host, to complete the metamorphosis into a juvenile mussel, what is a believed adaptation for dispersal (Kat, 1984; Graf & Cummings, 2006; Ferreira-Rodríguez et al., 2019) which probably appeared already before the Middle Jurassic when the modern lineages of Unionida already existed (Watters, 2001; Skawina, 2021). Adults are large (up to a dozen of cm in length) and live shallowly submerged in the deposits where they filter feed—in the 'siphoning position' of Archambault et al. (2014), Amyot & Downing (1991), Schwalb & Push (2007), Watters et al. (2001). On the other hand, juveniles, which commonly develop characteristic and species-specific shell sculpture (so-called umbonal sculpture, or umbonal rugae; Haas, 1969; Aldridge, 1999; Zieritz et al., 2015) live hidden within the bottom deposits.

Biology of early juveniles—what we know according to the scattered data

Larvae of unionoids are small [length from 60 to 360 µm, depending on the species; e.g., Wächtler et al. (2001)] and after absorbing nutrients from the body of their host they metamorphose and post-larval minute mussels, juveniles, escape from the host and sink to the bottom. If they find a suitable substratum, they burrow there and they undergo drastic morphological changes including the development of ciliated gills, which allows the transition from the juvenile's deposit feeding via the foot—to suspension feeding via gills. After reaching a size of about 3 cm—they emerge at the surface and begin adult life (Yeager et al., 1994; Wächtler et al., 2001; Schwalb & Ackerman, 2011; Schartum et al., 2017; Araujo et al., 2018; Irmscher & Vaughn, 2018). The post-larval or ‘early juvenile’ stage of life within sediments, as was introduced by Isely (1911) for mussels up to 15 mm in length (today the definition is understood as up to 3 cm in length) is currently believed to be the least known and one of the most vulnerable (Geist, 2010; Schartum et al., 2017; Patterson, 2018; Bílý et al., 2021). One of the reasons for the lack of basic biological knowledge is that minute juvenile unionoids are difficult to observe in the field e.g., (Lefevre & Curtis, 1910; Piechocki, 1969; Neves & Widlak, 1987; Amyot & Downing, 1991; Hastie & Cosgrove, 2002; Irmscher & Vaughn, 2018) and long were difficult to be grown in cultures to subadult age (Lefevre & Curtis, 1908; Dimock & Wright, 1993; Lima et al., 2012; Mair, 2018).

What we do know is that after their detachment from the fish host they are small (about 0.2–0.5 mm; Wächtler et al., 2001; Irmscher & Vaughn, 2018), may drift with the current (in streams) until they sink to the bottom of the stream/lake (where they can possibly exert some control over habitat selection; Schwalb & Ackerman, 2011; French & Ackerman, 2014; Irmscher & Vaughn, 2018) and immediately burrow themselves in well-aerated deposits (they are more mobile than older unionoids; Yeager et al., 1994; Wächtler et al., 2001; Kemble et al., 2020; Hyvärinen et al., 2021). Authors expect that general conditions are similar to those of the adults, but differ in microhabitats occupied (Neves & Widlak, 1987; Buddensiek et al., 1993; Strayer, 2008). They can live clustered (Neves & Widlak, 1987; Hastie & Cosgrove, 2002) or dispersed (Piechocki, 1969; Hastie & Cosgrove, 2002). More often they were found within fine deposits of small pools close to rivers' banks (Piechocki, 1969), in 'riffles, runs, and behind boulders' (but are rarer in pools) of the stream in pebbles, gravel, and sand (Neves & Widlak, 1987) or in ‘clean sand of lakes’ (James, 1985). Early juveniles of at least some species were observed to live byssally attached to submerged objects (Isely, 1911; Lavictoire et al., 2018). The depth to which early juveniles may burrow in nature is unknown, however, they were reported from a maximum depth up to 20 cm (Schwalb & Push, 2007), while in experiments the post–metamorphic early juveniles were recovered within the top 1 cm of sediment (Yeager et al., 1994), and more recently around 2–3 cm (Archambault et al., 2014; Kemble et al., 2020; Bílý et al., 2021). They are more commonly noticed when some tools [dredges, bottom scratch samplers, Bernatowicz’s grabs, or Günther’s sampler; see in Lewandowski & Kołodziejczyk (2014)] and intensive search are applied (Strayer, 1981; Amyot & Downing, 1991; Hastie & Cosgrove, 2002). Although field information is quite scarce, it is believed that they spend 2–3, up to 6 or even 20 years (Hochwald & Bauer, 1990; Amyot & Downing, 1991; Balfour & Smock, 1995; Strayer et al., 2004; Schwalb & Push, 2007; Simon et al., 2015) within the sediment in interstitial waters (hyporheal environment). There they pedal-feed [sensu Reid et al. (1992)] on unknown particles [possibly bacteria, detritus, different algae and diatom species (Gatenby et al., 1996; Jones et al., 2005; Barnhart, 2006)]. During this period, they undergo critical anatomical changes to establish the ability of suspension feeding.

The majority of the knowledge about these internal changes comes from laboratory cultures, designed for the conservation of a few dozen species at a maximum (e.g., Buddensiek, 1995; Gatenby et al., 1997; O'Beirn et al., 1998; Jones et al., 2005; Barnhart, 2006; Kovitvadhi et al., 2007; Schmidt & Vandré, 2010; Denic, 2018; with a detailed description of successful systems in Mair (2018)). In particular, Mair (2018) recently summarized that although mussel culturing has progressed and developed during the last two decades, there is no universal method for raising juvenile unionoids in the laboratory for conservation purposes. At the same time, quite common problems in culturing juvenile unionoids refer to their high mortality during the first months of life—in some cultures up to 100% (Lefevre & Curtis, 1908; Kovitvadhi et al., 2001; Jones et al., 2005; Eybe et al., 2013), but in more successful cultures the mortality dropped to 10–80% (Gatenby et al., 1996; Jones et al., 2005; Schmidt & Vandré, 2010; Eybe et al., 2013). This high mortality is interpreted as a result of diet or predation on juveniles (Kovitvadhi et al., 2001; Jones et al., 2005), or more recently as a result of possible food shortage—low filtering efficiency that accompanies organogenesis and the transition to filter–feeding [(Schartum et al., 2017; Araujo et al., 2018) but not confirmed by Lavictoire et al. (2018)]. The other described difficulties affecting survival of juveniles refer to their possible high sensitivity to environmental parameters in sediments [e.g., low oxygen concentration or high nitrate and ammonium levels; (Dimock & Wright, 1993; Mummert et al., 2003; Newton & Bartsch, 2007; Hyvärinen et al., 2022)]. This contributed to the idea to use early juveniles and their survival rates as a potentially useful biological indicator providing a measure of the quality of the streambed (Geist & Auerswald, 2007; Archambault et al., 2017; Kemble et al., 2020).

Juvenile shell ornamentation—biological role and evolution

In addition to the little-known biology, young unionoids (up to about 1 cm in length) typically bear pronounced surface protrusions [widely discussed by Zieritz et al. (2015)]. The sculpture’s functional role, while poorly understood, is often considered to be an adaptation to life within the sediment. It is believed to ease burying in the substrate and support anchoring in it (Stone et al., 1982; Watters, 1994; Hornbach et al., 2010), which represents one of the shell sculpture specializations typical for burrowing bivalves (Stanley, 1969; Seilacher, 1984; Savazzi & Yao, 1992). Nevertheless, the oldest Late Triassic unionoids did not have any prominent sculpture on the shell (Good, 1998; Skawina & Dzik, 2011) but it is known only since the Early or Middle Jurassic (Cai, 1986; Skawina & Dzik, 2011). The late emergence of prominent juvenile sculpture in the evolution of the unionoids may suggest that the discrepancy between the early juvenile and later stages of life characteristic of many extant unionoids emerged gradually during the Early Jurassic period. This interpretation of the unionoid phylogeny remains speculative until the actual functional meaning of the prominent juvenile (umbonal) sculpture is identified.

Aim of the experiment

The aim of this pilot experiment which was conducted as an observational study on a small number of animals was therefore to test two distinct methods of visualization of behavior within the deposits of unionoid juveniles of different, common in Poland (and in Europe) species. These species differ from the others by having prominent [Unio tumidus Philipson, 1788, U. pictorum (Linnaeus, 1758)] or delicate [Anodonta anatina (Linnaeus, 1758)] umbonal sculpture. My study also highlights the difficulties and feasibility of obtaining juvenile unionoids from the field in numbers that are sufficient for behavioral experiments. Limited access to the captive-breeding facilities at the time of this study on the one hand, together with the aim of tracing juveniles in the field in Central Europe decades after the last published attempt (Piechocki, 1969) on the other hand supported this endeavor. The scarce data from the literature cited above suggest that such minute individuals (preferred up to 10 mm, according to the common range of sculptured juvenile shell) have to be searched for with tools within the top 10 up to 20 cm layer of the bottom sediment. The limit of their penetration into sediments is presumably imposed by oxygen availability and contamination with toxic metabolites. They are also likely unable to penetrate the partly lithified clay that frequently underlies the sandy bottom of Polish rivers.

Methods

Fieldwork

The following techniques for collecting samples were applied, depending on the water depth and thickness of the oxygenated sediment layer:

  1. 1.

    Bottom scratch sampler—to a water depth of about 1 m, if the thickness of suitable sediment was small; side length 20 cm; (Lewandowski & Kołodziejczyk, 2014).

  2. 2.

    Spade—to a water depth of about 1 m, if the suitable sediment was more than 5 cm thick.

  3. 3.

    Günther sampler from the boat at water depths from about 1 m to about 5 m; sample catchment surface 272 cm2; (Lewandowski & Kołodziejczyk, 2014; Jurkiewicz-Karnkowska et al., 2017).

More than 30 localities were available to collect juvenile unionids, identified from literature data (e.g., Lewandowski, 1996 and personal information (advice) from K. Lewandowski and A. Piechocki, as well as my own assessment. They belonged to the public river–lake system of Krutynia (the Masurian Lakes, NE Poland; (Jakubik & Lewandowski, 2011). Sampling was performed in June 2009 and its aim was to search the promising deposits until success. Each sample was sieved through a 1 mm mesh geological sieve; animals, if found, were identified and placed in a container for transit. The length of each animal was measured after transfer to the laboratory. Fifteen live juvenile individuals less than 10 mm long (range 6–10 mm) were collected (two A. anatina (Linnaeus, 1758), one Unio pictorum (Linnaeus, 1758) and 12 U. tumidus Philipsson, 1788). All small mussels came from sandy deposits both in lakes and in rivers (nevertheless, the majority of the collected specimens came from rivers, commonly from shallow run water and close to river banks; they appeared dispersed—if any, not more than one individual per site was found); they were collected with a bottom scratch sampler (commonly in rivers) or spade (usually in lakes), and no juvenile bivalve was collected with the Günther sampler. Additionally, more than 20 juveniles measuring 11–30 mm (three A. anatina and more than 20 U. tumidus) were collected. All gathered individuals were kept alive in the water tank, however, only the 11 smallest individuals of all species (together with one 21 mm long U. tumidus) were selected for this observational study (thus requiring a small number of animals)—a pilot experiment (Table 1).

Table 1 The lengths of studied juvenile unionoids in every trial of experiment with their behavior—resting in an adult-like siphoning position or being completely submerged within deposits at the end of each trial (1–3 mm below the sediment surface, where possible to assess)
Full size table

All bivalves were sustained in a 60 l freshwater water tank with aerated water filtered with a carbon filter and were fed Scenedesmus sp. algae suspension (each 48 h) for 2 weeks before experiments. Each tank contained quartz sand deposits of about 1–2 cm in depth (sieved with a geological sieve, 0.25–1 mm fraction, following the field observation of common strata). The quartz sand was washed and sterilized after purchasing from a sand mine in the Vistula River (in Warsaw), in which adult unionoids are observed abundantly but where juveniles haven’t been gathered (pers. obs.).

X-ray method

To trace the dynamics of vertical behavior of juvenile unionoids within sediments, I applied an X-ray method that requires laminated deposits of different density contrast (following idea of Gingras et al., 2008). After the bivalve burrows deeply, it should leave a pathway of bioturbated, mixed sands of two kinds, visible on X-ray images. First, in an attempt to identify the best contrast of deposit lamination in X-ray pictures, several successions of laminations were prepared containing layers of quartz sand, sand with quartz grains stained with iron minerals, and aragonite sand (commercially prepared for marine coral–reef aquaria and purchased from the zoological shop; washed and sterilized prior to use). Glass test tubes (19.0 cm high and 1.5 cm in diameter) were filled with interbedded laminae of fine–grained sands (Fig. 1). The laminations were obtained by filling the tubes with water and sprinkling sand into them by hand. The thickness of a single layer was about 5 mm. To examine the suitability of the method for X-ray imaging empty juvenile shells of U. tumidus or A. anatina of similar size (about 8 mm) were placed in the middle of each test tube. The tubes were X-ray imaged at the Philips Duo Diagnost (SN: 7001487) diagnostic device in the Department of Radiology, Military Institute of Medicine, Warsaw. Files were produced as DICOM files. All shells were at least noticeable as blackish areas within lighter sands (Fig. 1c), and the best contrast between layers was achieved by using laminas of quartz and aragonite sands, thus this lamination was used to fill the glass aquariums.

Fig. 1
figure 1

Test tubes prepared for X-ray examination (a) and their X-ray picture (b). c Magnification of the selected area of b showing blackish areas—the position of the bivalved shell (empty space, restricted by the shell, within more dense—bright—deposits). Stars (a) and arrows (b) indicate the position of empty juvenile shells

Full size image

Two flat, glass aquariums (dimensions 25.0 × 25.0 × 1.5 cm) were filled with 5 mm thick layers of fine–grained sand (alternately quartz layer and aragonite layer), following the procedure first described in Gingras et al. (2008), see above. Both aquariums were aerated, kept in a room temperature, and in a natural light-darkness regime of a long day (18 h light: 8 h darkness). Juvenile naiads were then placed in the aquariums on the surface of the top quartz sand layer. The first aquarium accommodated one 8 mm long individual of U. tumidus while the second one contained two 9 mm and 14 mm long individuals of A. anatina. Bivalves were fed with Scenedesmus sp. suspension every 48 h (around 25 ml of thoroughly mixed algae suspension each time, added directly to the aquarium water). At variable time steps—1, 25, 72 h and after a week, X-ray images were produced (Fig. 2). The sequences of X-ray images were then used to determine the amount of sediment disruption and the track of the animal.

Fig. 2
figure 2

The flat aquaria for X-ray examination. Three bivalves are visible in the top layer of sediment or indicated by blackish areas in bright sand (arrowed). a Experimental set after starting the first hour of the experiment; bivalves from this trial: U. tumidus 8 mm (b) and one of A. anatina (9 mm) (c); X-ray pictures: flat aquarium after a week (d) U. tumidus; magnification of view of top layers with bivalves in flat aquariums: after 25 h (e) U. tumidus, and g two A. anatina (9 and 14 mm); after 72 h: f U. tumidus and h two A. anatina

Full size image

Transparent sand

For direct observations and video–recording of the mussels’ behavior within the sediments, ‘transparent sand’ consisting of silica gel was tested. While the material is sold as a desiccant comprised of solid white beads 1–3 mm in diameter (product no. 908250460, POCH S.A.), it becomes immediately quite transparent after soaking in water. No harmful effects were observed on the animals’ behavior during the experiment. Because the grains in the packages were often crushed during transit, the dry silica gel was first sieved to exclude fractions smaller than 0.25 mm.

Two other flat glass aquariums of the same dimensions (25.0 × 25.0 × 1.5 cm) were filled with pre-washed silica gel (approximately 15 cm deep layer) of and used for observations and video recording of juvenile behavior. In the first aquarium two juvenile mussel individuals were introduced—U. tumidus and U. pictorum, both 8 mm long; in the second one—three individuals of A. anatina measuring 6 mm, 9 mm [the individual from the X-ray experiment was re-used here], and 14 mm in length. The two flat water tanks were continuously recorded on video for 4 and 5 days, respectively, with the Sony DC-RSR 37E camera. Bivalves were fed with Scenedesmus sp. suspension (around 25 ml, feeding procedure as described above) every 48 h, and aquariums were aerated, kept in a room temperature, and in constant light (due to the limitations of the recording possibilities of the camera). In addition, video–recording was performed for another two individuals of U. tumidus (8 and 10 mm long), after finishing the first experiment (in the same aquarium and after removing previous Unio individuals, with the same observing regime).

Shallow aquarium

To trace the horizontal movements and accompanying behavior of the bivalves light photographs of the surface of deposits with U. tumidus measuring 21 mm, 10 mm, and 8 mm in length were taken in a shallow aquarium (11 × 11 cm and 15 cm high) with standardized as above quartz sand (1.5 to 2.0 cm deep to not exceed the depth of the sand already verified as ensuring good quality of the planned X-ray images). Bivalves were fed with Scenedesmus algae suspension every 48 h, aquarium was aerated, kept in a room temperature and a natural light-darkness regime of long day (18 h light: 8 h darkness). The camera used was Olympus E-510. The aquarium was light photographed from the top two times—the next day after setting mussels and a week later. Then after a day, X-ray pictures were taken from the top to show the number and position of bivalves in the sediment.

Results

X-ray method

Both kinds of sediment (quartz and aragonite) for this method did not vary in grain size in nature—at least quartz sand, as a top layer, was supportive and penetrable for juveniles (they were able to settle within the top layer). Nevertheless, no—or only minimal sediment disruption was recorded and no animal’s trace was recorded. All animals rested in the top 0.5 cm layer of quartz sand after obtaining the ‘adult’ siphoning position, without an attempt to burrow into deeper layers during the experiment.

The activity of U. tumidus within the sediment demonstrated by X-ray photography was very low (Fig. 2, Online Resource 1). The naiad was passively lying on the sediment surface for the first hour. During the next few hours, it set itself in the siphoning position typical for adult unionoids: the anterior part of the body was buried in the sediment, the posterior exposed above its surface (Fig. 2e). This position was recorded after 25 h, and only after 72 h the images showed a small horizontal shift of the mussel (2 cm) that was now almost completely covered with sand. All movements were horizontal and near the surface of the deposit, within the first lamina of quartz sand that was occasionally mixed with the aragonite grains underlying it (left area of Fig. 2f, compared to Fig. 2e). Both specimens of A. anatina attained the ‘adult’ siphoning orientation during the first 25 h of the experiment and there were no changes in their position until the end (Fig. 2g, h).

Transparent sand

The silica-gel sediment was sufficiently stable, supportive, and penetrable for the bivalves, and was quite transparent (larger crystals ensured more transparency than smaller ones), allowing direct observations of mussel behavior within it. The observed behavior was similar between prominently sculptured juvenile Unio species and delicately sculptured A. anatina. Individuals from all three species behaved in a similar way (Fig. 3), initiating locomotory movements in minutes after their introduction to the aquarium, and finishing movements when settled in an adult position. Only the smallest A. anatina differed from them in both timespans to begin locomotory movements (longer) and the way of performing these movements.

Fig. 3
figure 3

Freeze pictures from the video recording of the behavior of naiads in flat aquaria filled with silica gel. a Two specimens of A. anatina (14 mm and 9 mm). b Unio tumidus (left) and U. pictorum (right), both 8 mm long. c, d U. pictorum while digging deep into the sediment

Full size image

The 8 mm long individual of U. tumidus needed 17 min to start locomotion after the initial placement in the aquarium; it then settled in the ‘adult’ position after about 15 min of movement. U. pictorum (8 mm long) started to move after 3 min and found the appropriate place to settle after about 2 h of mobility. The largest individual of A. anatina (14 mm long) required almost 2 min to start moving and settled after 21 min of moving while the 9 mm individual, began locomotory movements after 4 min and settled in an ‘adult’ position after 1 min. The smallest, 6 mm long individual stayed motionless for 9.5 h, then moved to settle down for about 1 h, in an unusual way (Fig. 4).

Fig. 4
figure 4

Freeze pictures of the unusual movements of minute 6 mm long A. anatina (line serves as reference mark). Note the position of the pedal gape during the movement (direction of movement from left to the right—from a to d)

Full size image

An 8 mm long individual of U. pictorum was observed to settle under the sediment surface (Fig. 3d) and a similar behavior was observed in all 3 individuals of U. tumidus (Fig. 5; Table 1). In each case, the sediment covering the animals was a few millimeters deep. No deeper burrows were observed, but the animals treated this position as suitable for settlement.

Fig. 5
figure 5

Freeze pictures of the movement of U. tumidus. a 10 mm long individual settled down (left one) while 8 mm long one (the right specimen) was in motion. b Settlement of the 8 mm long individual. c Magnification of b, showing the distance from the posterior end of the bivalve to the sediment surface. d 10 mm long individual digs deep into the sediment

Full size image

Shallow aquarium

The behavior of the three individuals of U. tumidus kept in the shallow aquarium differed by size (corresponding to age), but the sample size was too small to draw general conclusions. The larger bivalve was visible from above, while the smaller ones were completely covered with sand (Fig. 6; Table 1). Although the depth of the sand did not exceed the length of the largest bivalve, observations suggested that it would be possible for it to burrow itself completely inside the deposits because the position of the bivalve within the deposits is not vertical. X-ray photography (Fig. 6d) shows the position of all three animals.

Fig. 6
figure 6

The juvenile U. tumidus kept in a shallow aquarium. a The specimens used in the experiment. b All the specimens are visible on the surface the next day after setting into the aquarium. c The surface of the sand with one specimen visible and the others completely submerged in the sediment a week later. d X-ray picture showing localization of all three individuals

Full size image

Discussion

Despite more than a century of research, the hidden hyporheal period of the lives of juvenile mussels remains underexplored because early juveniles are difficult to trace in nature, as well as maintain and grow in artificial cultures (Lefevre & Curtis, 1908, 1910; Isely, 1911, Piechocki, 1969; Hastie & Cosgrove, 2002; Mair, 2018). Researchers most commonly refer to clear aerated sand and gravel in both the rivers and lakes as a habitat of juveniles (Piechocki, 1969; James, 1985; Neves & Widlak, 1987; Hastie & Cosgrove, 2002). In this study, the attempt to gather minute unionoids resulted in 15 individuals (< 10 mm in length) and all the collected juveniles of three species came from clear, aerated, sandy habitats in shallow water, both in rivers (majority of mussels) and lakes (several individuals). The better success in the catchment of riverine juveniles was possibly a result of a shallow layer of aerated sand (about 5–10 cm) with a partially lithified clay layer underneath, which was likely impenetrable to mussels. Therefore, it was easier to search the entire available mussels’ habitat. The most useful tool was a bottom scratch sampler, supported by a spade, while the Günther sampler appeared less useful likely because this tool sampled only a few top centimeters of the deposits with quite small square dimensions, and a thorough search was unlikely (insufficient and inefficient sampling). The importance of sampling efficiency and sufficiency is underlined by Neves & Widlak (1987) and Hastie et al. (2010). Additionally, although Neves & Widlak (1987) and Newton et al. (2008) described juveniles as clumped in distribution, my study does not support this observation in line with previous studies from mussel populations in Scotland (Hastie & Cosgrove, 2002) and Poland (Piechocki, 1969).

Gingras et al. (2008) proposed time–stepped X-ray images of laminated deposits in which they determined the amount of sediment disruption made by marine invertebrates. In their method, the interbedded quartz with heavy mineral sand gave a sufficient result. I supplemented this method with aragonite sand, which is easy and cheap to provide, and gave a clear visual contrast when interbedded with quartz grains. The mobility of marine invertebrates reported by Gingras et al. (2008) was related to the mode of food gathering—suspension feeders disrupted the sediment by burrowing 10 to 100 times less than deposit–feeding taxa. Scientific interest in the diet and the way the food is gathered by juvenile unionoids has grown since an influential study by Yeager et al. (1994), which described the feeding and burrowing behavior of minute Villosa iris (Lea, 1829) individuals and documented their pedal–sweep, pedal–locomotory and interstitial suspension feeding behaviors. Araujo et al. (2018) specified that juveniles first feed with their ciliated foot and then filter suspension by developing gills, but it is not yet determined what type of food is ingested during the pedal– and filter–feeding periods of juvenile life.

Although juvenile unionoids are regarded as both deposit and suspension feeders (Yeager et al., 1994), my pilot study shows that they were not burrowing deeply and that there was no deep reworking of any of the deposits. All bivalves rested in an adult-like siphoning position after a short period of mobility and were covered with quartz sand within the top layer during the X-ray experiment (however, a slightly reworked area on the previous border of the first quartz and aragonite layers is visible after 72 h and a week of trial, in the left area of the aquarium of U. tumidus, Fig. 2d, f) and were placed a little deeper, but still within the top 1 cm,—in silica gel (Figs. 2, 3 and 5). Although a number of studies indicate that juveniles are supposed to burrow at least to a depth of several cm (e.g., Neves & Widlak, 1987; Wächtler et al., 2001; Schwalb & Push, 2007), similar results were previously obtained in laboratory experiments conducted by Yeager et al. (1994) (with early V. iris juveniles that were recovered from the top 1 cm of sediment), Kemble et al. (2020) (commonly up to 3.4 mm only, and not deeper than 5.1 mm), or Archambault et al. (2014) (not deeper than 2.5 cm). However, as expected, small (8–10 mm long) U. tumidus were observed to burrow deeper in the sediment than larger individuals (Figs. 3, 5 and 6; Table 1). Juveniles of relatively advanced age may maintain the typical adult siphoning orientation (Figs. 3, 6; Table 1). The shallow burrowing depth may be the result of unnatural experimental conditions (e.g., the sterilized quartz sand did not contain detritus and both aragonite sand and beads of silica gel were not natural to them), or of biological factors as for example experimental animals might have been collected for experiments after reaching the age at which they actually begin the post-juvenile life mode in close proximity to epibenthic environments. Hyvärinen et al. (2021) observed that the size of the grains within the deposit is important for the early juveniles of Margaritifera margaritifera—they burrowed well when grains were at least 0.25 mm (but much worse, when smaller), and the best, when grain size was above 0.5 mm in diameter, which is similar to the grain size used in my experiment. Nevertheless, Kemble et al. (2020) did not consider grain size an important determinant of the depth of burrowing of juveniles. Schwalb & Push (2007) suggest that burrowing behavior may depend on a variety of environmental factors. As laboratory conditions by definition exclude many of them, the juvenile mussels might lack the right environmental trigger for deep burrowing. Similarly, the supply of washed and sterilized deposits might have prevented bivalves from their presumed search for food in the interstitial water (Gatenby et al., 1996; Archambault et al., 2014). Feeding on the sediment may provide a significant portion of the total energy for locomotion (Vaughn & Hakenkamp, 2001).

Sparks & Strayer (1998) described a ‘stress surface behavior’ when a bivalve does not burrow itself in the deposits as a result of a stress reaction to unfavorable conditions. My results, however, do not support this explanation as the experimental bivalves did burrow quickly, albeit not deeply. Finally, Archambault et al. (2014) considered that the burrowing depth might be linked to the season because unionoids tend to burrow deeper in winter than in summer (Amyot & Downing, 1997; Perles et al., 2003; Saarinen & Taskinen, 2003; Schwalb & Push, 2007), which cannot be excluded in the above trial as the experiment was conducted during the summer months.

If the anatomical development of gills and the subsequent transition from pedal feeding to filter feeding are sufficient to trigger a shift to an adult mode of life outside sediments, the emergence of juveniles to the surface of the deposits could occur earlier. This idea may be weakened by the recent observations of Schartum et al. (2017) who showed that the critical size for a juvenile M. margaritifera to start switching from juvenile pedal feeding to adult filter feeding is about 2.2 mm in length and is related to size only, not age. Additionally, M. margaritifera that exceeds 4.5 mm in size must filter feed because pedal feeding is not sufficient. At the same time, a 1.5 cm long juvenile examined by the authors had underdeveloped gills. Mair (2018) noticed that, in general, mortality in cultures usually decreases when the juvenile mussels exceed 1 mm in length. This is consistent with the timing of the transition from pedal feeding to filter feeding recognized in juveniles of M. margaritifera and U. mancus Lamarck, 1819 (Araujo et al., 2018). A proper transition from pedal feeding to gill feeding during juvenile ontogenesis has recently been identified as essential for survival (Schartum et al., 2017; Araujo et al., 2018) but not by Lavictoire et al. (2018). Unfortunately, the timing of this organogenesis has been studied in a small number of species, and it varies between species, e.g., in Hyriopsis (Limnoscapha) myersiana (Lea, 1856) the development of gills is quicker than in Anodonta (Kovitvadhi et al., 2007). In this context, a single observation of an unusual behavior in a 6 mm long A. anatina (Fig. 4; Online Resource 2) could be interpreted as related to food intake behavior [e.g., pedal–locomotory feeding; (Yeager et al., 1994)] even though the bivalve was not inside the sediment, especially when Kaestner (1967) describes the development of the outer demibranchs of Anodonta which is not complete until its shell size reaches 3 to 5.7 mm. Normally, any locomotory movement of an adult naiad requires (1) a slight opening of the shell valves with protrusion of the foot into substrate until fully (or almost fully) extended, (2) anchoring the tip of the foot by dilation of its distal end, and (3) contraction of the pedal retractor that results in pulling the shell towards the anchored foot. After such movement, the shell is usually slightly tilted, its anterior end being oriented downward, and the ventral margin is horizontal (see, e.g., Figs. 2, 3a—the larger individual). The behavior of the smallest A. anatina (6 mm) was unusual in this respect. Although it was not buried in the sediment (as would be predicted for a juvenile), it moved in a manner different from that of the larger juvenile bivalves. Pulling off its foot rotated the shell to an upright position, with the anterior end towards the bottom, but the ventral part of the shell was oriented in the direction of movement (Fig. 4; Online Resource 2). The sculpture on the juvenile shell of unionoids is a feature that—with the hyporheal period of early juvenile life—may represent an adaptation to life inside the sediment (Seilacher, 1984; Savazzi, 1991). There are many possible reasons for evolving and supporting the burrowing behavior of juveniles including protection from larger predators and possible parasites or competitors [summarized by (Strayer, 2008), but also exampled by (Nichols & Wilcox, 1997; Taskinen & Saarinen, 2006; Schwalb & Push, 2007)]. However, the design of the current experiment does not permit further discussion of the functional aspects of juvenile sculpture based on empirical evidence. Due to the difficulties in obtaining minute juvenile unionoids from the field (resulting from their biology) in sufficient numbers to achieve statistical significance, the study shows potential for continuation in the future, possibly supplemented by incorporating juveniles from the captive-breeding facilities (like in Černá et al., 2017).

Conclusions

The rapid and global decline of unionoid freshwater mussels, as well as many outstanding questions about their biology during one of the most vulnerable and little-known stages of their life, led me to design two new experimental methods for visualizing the behavior of juvenile unionoids within sediments: (1) time-stepped X-ray images of bivalve traces within laminated deposits of quartz–aragonite sand and (2) silica gel serving as ‘transparent sand’ for direct video recording of the behavior within sediments. The methods were cheap and easy to set up. The deposits were stable, supportive, and penetrable (no signs of harmful effects were observed during the experiments) and were readily used for burrowing by juvenile unionoids. Although the experiment was conducted as an observational study on a small number of animals to test the methods, the findings may be informative for more substantial research in the future.

All tested in this study juvenile mussels settled in the top 1 cm layer of deposits with no attempts to burrow deeper in both methods. More data are needed, possibly with the use of younger stages of mussels (e.g., from captive-breeding facilities), to give a more conclusive answer if such behavior was the result of biological factors (e.g., they were collected for the experiment after they reached the post-juvenile stage) or the unnatural experimental conditions. However, consistent with my predictions, smaller mussels were observed to burrow deeper and were more thoroughly covered by the deposits than larger individuals.