A peer-reviewed open-access journal @) NeoBiota Advancing research on alien species and biological invasions NeoBiota 94: 243-259 (2024) DOI: 10.3897/neobiota.94.119622 Research Article Dead or alive: the effect of shells and living individuals of Sinanodonta woodiana (Lea, 1834) on habitat selection and behaviour of European unionid bivalves Kamil Wisniewski'®, Daniel Szarmach'®, Jarostaw Kobak'®, Tomasz Kakareko2®, Lukasz Jermacz2®, Matgorzata Poznariska-Kakareko'® 1 Nicolaus Copernicus University in Torun, Faculty of Biological and Veterinary Sciences, Department of Invertebrate Zoology and Parasitology, Torun, Poland 2 Nicolaus Copernicus University in Torun, Faculty of Biological and Veterinary Sciences, Department of Ecology and Biogeography, Torun, Poland Corresponding author: Kamil Wisniewski (kam.wis@doktorant.umk.pl) OPEN Qaceess Academic editor: Belinda Gallardo Received: 26 January 2024 Accepted: 12 June 2024 Published: 7 August 2024 Citation: Wisniewski K, Szarmach D, Kobak J, Kakareko T, Jermacz t, Poznanska-Kakareko M (2024) Dead or alive: the effect of shells and living individuals of Sinanodonta woodiana (Lea, 1834) on habitat selection and behaviour of European unionid bivalves. NeoBiota 94: 243-259. https://doi.org/10.3897/ neobiota.94.119622 Copyright: © Kamil Wisniewski et al. This is an open access article distributed under terms of the Creative Commons Attribution License (Attribution 4.0 International - CC BY 4.0). Abstract 1. Ecosystem engineering freshwater bivalves, burrowing in the substratum and accumulating shell deposits, transform bottom habitats. Especially the invasive Asian bivalve Sinanodonta woodiana (SW), due to its rapid growth, large size, and high fecundity, can affect benthic communities. Here, we determined its effect on habitat selection and behaviour of endangered native bivalves, Anodonta cygnea and Unio tumidus. 2. We conducted laboratory preference assays (Experiment 1: choice between two substrata) exposing the native bivalves to pure sand (control), shells (several densities on the sand surface or burrowed), or living SW. Then, we tested their locomotion and burrowing (Experiment 2) on pure sand and substrata contaminated with shells or living SW. 3. In Experiment 1, native bivalves avoided shells, but not living SW. Burrowed and larger shells were avoided compared with those on the surface and smaller ones, respectively. 4, In Experiment 2, U. tumidus exposed to SW delayed activity initiation (in response to living bi- valves), increased locomotion (living bivalves, surface shells), and reduced burrowing depth (living bivalves, all shells). Anodonta cygnea exposed to SW reduced locomotion speed (living bivalves, shells), and reduced burrowing duration (burrowed shells) and depth (living bivalves, burrowed shells). 5. SW (especially shell beds) constitutes another emerging threat to native bivalves, impairing their burrowing and inducting active avoidance. As SW expands its distribution with climate warming, the range and strength of its impact is likely to increase, reducing the area available to native bivalves, exposing them to environmental dangers (due to burrowing limitation) and deteriorating physical condition (energetic resources used for excessive locomotion). Key words: Behaviour, biological invasions, Bivalvia, ecosystem engineers, habitat selection, inter- specific interactions, species displacement, unionid mussels 243 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae Introduction Bivalves of the Unionidae family are freshwater bottom dwellers of limited mobili- ty (Curley et al. 2021). Through filter-feeding, bivalves can considerably modulate the availability of resources for other organisms by transferring suspended parti- cles to the bottom sediments (Boeker et al. 2016; Pouil et al. 2021). Both living bivalves and their empty shells, accumulating in the environment long after the animal death, constitute unique hard structures affecting community functioning (Gutiérrez et al. 2003). Therefore, bivalves are considered to be ecosystem engi- neers and their extinction, overpopulation or changes in their taxonomic composi- tion lead to habitat modifications with cascading effects on the aquatic community and ecosystem services provided by these animals (Vaughn 2018). Freshwater bivalves are threatened globally by human impact, including climate change and pressure from non-native species (Strayer 2008, Blackburn et al. 2014; Bacher et al. 2018). The number of alien bivalve species and sizes of their popu- lations have been growing for several decades (Seebens et al. 2021; Latombe et al. 2022). Accordingly, the pressure from non-native bivalves is considered a major threat to their native counterparts (Mack et al. 2000; PySek et al. 2010), especially due to their ecosystem engineering properties and resulting habitat transforma- tions (Bddis et al. 2014a; Douda et al. 2024). The Chinese pond mussels of the genus Simanodonta are unionid bivalves native to Eastern Asia, but invasive in other parts of the world. Recent genetic studies have shown that invasive lineages belong to three species: (i) S. woodiana (Lea, 1834), the “temperate invasive” lineage, native to southern China and invasive in Europe, as well as in western and central Asia, observed probably in Africa (find- ing needs genetic confirmation (Bensadd-Bendjedid et al. 2023)); (ii) S. pacifica (Heude, 1878) the “tropical invasive” lineage, whose native area is Taiwan and eastern China, whereas it is invasive in North America, southern Asia and Iraq, and (iii) S. dauta (Martens, 1877), originating from Japan, the Korean Peninsula and eastern Russia, and invading central and southern Asia (Douda et al. 2024). In Europe, only S. woodiana (“temperate invasive” lineage) occurs, likely originating from a single introduction event (Koneény et al. 2018). Therefore, we will focus on this species in the current study. In colder regions (such as central and eastern Europe), its spread was initially limited to artificially heated waters (Urbaniska et al. 2012), but, over the past two decades, it has accelerated and extended to habitats of natural thermal regime (Bogan et al. 2011; Bolotov et al. 2016; Lopes-Lima et al. 2017; Bespalaya et al. 2018; Kondakov et al. 2018, 2020; Koneény et al. 2018). Substratum preferences of S. woodiana overlap with those of the native European Unionidae (Poznariska-Kakareko et al. 2021) indicating a high risk of competitive tensions between them (Douda and Cadkova 2018). This invasive bivalve exhib- its a number of competitive advantages over the native Unionidae, including the higher rate of host infection by its parasitic glochidium larvae, faster development and growth rate (Douda et al. 2012; Huber and Geist 2019) and higher fecundity (Labecka and Domagala 2018; Labecka and Czarnoleski 2019). Another potential mechanism of the impact of S. woodiana on native Unioni- dae can be the transformation of the bottom by living individuals and shell beds formed after the bivalve death (Bédis et al. 2014a; Nakano 2023). Sinanodonta woodiana is a large (up to 26 cm) and fast-growing species (Urbariska et al. 2019) NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 244 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae reaching high densities. In a Polish lake, densities of 68 ind. m’ and 27.9 kg m* were observed (Kraszewski and Zdanowski 2007), which is the highest density of this species reported for Europe. In other European countries, the density ranges from a few ind. m®* in Ukraine (Yermoshyna and Pavliuchenko 2021) to c.a. 50 ind. m* in Hungary and Italy (Benké-Kiss et al. 2013; Kamburska et al. 2013). Over time, shells accumulate on the bottom surface and in the sediments, outnumbering living individuals and forming a layer significantly changing the substratum quality. More than 280 ind. m” (counting both valves as one indi- vidual) were noted by Béddis et al. (2014a). Shells reduce the near-bottom cur- rent velocity, limit the access of light to the bottom, and increase microhabitat heterogeneity (Gutiérrez et al. 2003). Moreover, shells (especially large ones) can create physical barriers limiting bivalve movement and burrowing, and thus degrading the living conditions for these organisms. It is likely that the effect of shell beds formed by S. woodiana will be stronger than that of shells origi- nating from the native species present in the environment before the invasion, due to the shorter lifespan of S. woodiana. It can live up to a maximum of 12 years (Spyra et al. 2012), compared to the maximum lifespan of 37 and 21 years exhibited by native Anodonta cygnea (Linnaeus, 1758) and Unio tumidus Philips- son, 1788, respectively (Aldridge 1999). This results in a faster accumulation rate of S. woodiana shells on the bottom. Moreover, its shells are larger, thus consti- tuting larger and heavier physical obstacles in the substratum. Knowledge of the responses of the native bivalves to the presence of S. wood- iana will help understand the mechanisms and magnitude of its impact, as well as develop methods of dealing with this new threat. The aim of our study was to determine mechanical effects of substratum contamination with living individuals and shells of S. woodiana on behaviour (habitat selection, locomotion and bur- rowing) of two native European unionid bivalves: A. cygnea and U. tumidus. Their numbers are constantly decreasing worldwide (Lopes-Lima et al. 2017), and they are protected by law in several countries (Van Damme 2011; Lopes-Lima 2014). These species were selected due to their reported coexistence with S. woodiana (Lajtner and Crnéan 2011; Beran 2019) and similar habitat preferences (Poznaris- ka-Kakareko et al. 2021). We hypothesized as follows: (1) native bivalves would avoid substrata contaminated with S. woodiana. (2) The adverse effect of shell beds on bivalve preferences would result from deteriorated burrowing and/or lo- comotion. Alternatively, increased locomotion might indicate active avoidance of the substratum contaminated by S. woodiana. (3) Empty shell beds would affect native bivalves to a greater extent than living S. woodiana. This might be due to (i) variable shell positions in the sediments (horizontally or vertically, on the surface or burrowed) compared to always vertically burrowed living bivalves (see Suppl. material 1: fig. S1), or (ii) the presence of sharp shell edges irritating the foot of moving bivalves. Options (i) or (ii) would be supported by stronger unionid responses to shells present on the sediment surface or burrowed, respectively. (4) The effect of S. woodiana shells on native bivalves would differ from that of native shell beds. A presumably stronger effect of the invader (compared to the shells of native bivalves) would result from either interspecific differences in shell structure (resulting in different unionid responses to shells of various species presented at the same sizes and densities) or the larger size of S. woodiana shells (resulting in stronger responses of unionids to larger shells). NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 245 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae Materials and methods Bivalve collection in the field Anodonta cygnea, U. tumidus and S. woodiana (shells and living individuals) were collected in early autumn from the sandy/muddy bottom (depth: 1.5—2.5 m) from the Wloclawski Reservoir on the River Vistula, Central Poland (52°37'04'"N, 19°19'42"E) by scuba divers. This site represents a natural thermal regime for cen- tral Europe, and has been recently invaded by S. woodiana (Cichy et al. 2016; Douda et al. 2024). Currently, all the study species co-exist at the location and the invader is constantly increasing its abundance and range (personal observation). We obtained S. woodiana shells from freshly killed individuals (on the day of col- lection), while native Unionidae shells were collected as fresh shells (uncorroded, undamaged) lying on the bottom of the reservoir (to avoid killing the native spe- cies). The collected bivalves were transported to the laboratory in buckets with substratum and water from the reservoir and tested after two weeks of acclimation. Stocking conditions Living bivalves (each species separately) and empty shells were kept in 350-L stock tanks (20-30 individuals per tank) equipped with internal filters and aeration sys- tems, with the bottom covered by a few cm deep layer of sand taken from the collection site. The stock/experimental room was equipped with a photoperiod sys- tem (light/dark cycle: 12:12 h) imitating the natural day-night cycle, and air-con- ditioning which kept the water temperature in the tanks at the level similar to that observed in the reservoir during bivalve collection. We checked the water quality in the stock and experimental tanks using a multimeter Multi340i (WITW GmbH, Weilheim, Germany). The water parameters were within the following ranges: ox- ygen content: 7.37—7.77 mg ml' (82.9-87.2%); temperature: 18.4—20.1 °C; pH: 8.01—8.67; conductivity: 643-827 uS cm. The bivalves were fed twice a week with a suspension of dried Chlorella algae (“Chlorella super alga”, Meridian compa- ny, Poland) in a concentration of 5 mg L (Douda and Cadkov4 2018). Experiment 1: Habitat selection Tests were conducted in 30 x 30 x 30 cm tanks divided into halves (Suppl. material 1: figs S1, S2). Each half was filled with a different substratum (see below) to a depth of 10 cm. Then, the tank was filled with conditioned (settled and aerated for at least 48 h) tap water (a 10-cm layer above the substratum surface). One bivalve individual was introduced in the central part of the tank with its ventral side down and anteroposterior axis parallel to the border between the substrata. After 24 h, the location of the tested individual was checked (choosing one of the two substra- ta). Each configuration of substratum types was repeated 30 times per species. We used the following substratum types: (1) Sand (grain diameter range: 0.2—1.4 mm; median: 0.63 mm) obtained from the bivalve collection site. This material was earlier found to be preferred by all the species tested (Poznariska-Kakareko et al. 2021). The pure sand was used as a control. The same sand type was contaminated with S. woodiana to create other substratum types. NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 246 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae (2) Empty S. woodiana shells, small (mean length: 7.6 cm) or large (12.6 cm) (Table 1), composed of two valves connected naturally by the ligamentum. The shells were burrowed under the surface (covered with sand) or laid on the sur- face of the sandy substratum. We used these two shell position variants because a bivalve usually dies on the surface and, after some time, due to hydrodynamics and sedimentation, its shell becomes burrowed. ‘This is especially visible in the case of mass mortalities, when large quantities of empty shells cover the surface of the bottom (Bodis et al. 2014b). Shell arrangement in/on the substratum (ly- ing on their side or put vertically with their ventral surface down) was random (Suppl. material 1: fig. S1) to reflect their arrangement in the field. Shells were randomly put on the sand within the tank half they were assigned to (Suppl. material 1: fig. S1). Then, those assigned to the burrowed variant were gently and thoroughly covered with sand to fill all the spaces between them. (3 SS Living individuals of S. woodiana (mean length of 11.6 cm, corresponding to the large shells; Table 1) immobilized by adhesive tape applied to the front of the shell (to prevent their relocation) and burrowed in the sandy bottom at 75% of their length (Suppl. material 1: fig. S4). The immobilized bivalves could not extend their foot and move, which was necessary to keep them in their positions within the assigned half of the tank (Suppl. material 1: fig. S4). However, they could partially open their valves, pull out the siphons, filter water and breathe. All bivalves and shells were thoroughly rinsed with water before use and biofilm and adhering debris were scrubbed from their surfaces. The sand was rinsed and dried in a laboratory dryer (SLW 115 STD Multiserw-Morek, Poland) at 60 °C for 6 h before use to eliminate any organisms that could potentially affect the results of the experiment. It should be noted that the size defined as large in our study is not of the maximum size of S. woodiana (26 cm, Urbariska et al. 2019). These, however, can be generally collected from warmer waters, whereas we used the size range commonly available at the collection site of the thermal regime natural for central Europe. First, we checked unionid selectivity between the pure sand and various shell den- sities (small or large, on the surface or burrowed). We started the experiment with a density of 133 ind. m® (6 shells per tank, two valves counted as one individual), i.e. twice as much as the maximum field density observed in heated waters. Then, we continued with the lower (67 ind. m”, 3 shells per tank) or higher (200 ind. m’, 9 shells per tank), depending on the presence or absence of a significant reaction to the initial density, respectively. This allowed us to determine the minimum effective density capable of influencing bivalve behaviour. We also confronted the pure sand with living S. woodiana at a density of 133 ind. m*. We did not use higher densi- ties of living S. woodiana, as they would have been unrealistic given the maximum density reported in the wild (Kraszewski and Zdanowski 2007). Moreover, we confronted the following: (i) burrowed shells vs. shells present on the sediment surface (using small shells at a density of 200 ind. m”) to check if shell position makes a difference, (ii) living S. woodiana vs. large burrowed shells (133 ind. m”) and (iii) burrowed small vs. burrowed large shells (200 vs. 133 ind. m”, corresponding to the same total volumes occupied by shells of the two sizes) to check whether bivalves respond differently to shell beds composed of shells of differ- ent sizes, (iv) native unionid shells vs. pure sand, (v) native unionid shells vs. small S. woodiana shells, to check if unionid responses to shells depend on shell origin. NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 247 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae Table 1. Total length of bivalves and shells [cm]. Mean SD Range A, cygnea 10.4 0.83 9.0-13.0 U. tumidus Vises 0.60 6.5-8.5 Native bivalve shells* 7.0 0.88 5.5-9.0 S. woodiana living individuals 11.6 0.84 10.5-13.0 S. woodiana small shells 7.6 1.14 5.5-10.0 S. woodiana \arge shells 12.6 1.24 10.5-14.5 SD - standard deviation, * - U. tumidus with small admixture of U. pictorum and A. anatina. Native shell beds were composed mostly of U. tumidus shells with a small ad- mixture of U. pictorum and A. anatina (as they occurred in the field). They were of a size considered in the current study to be small (Table 1) and burrowed in sand (in this form they triggered stronger responses in earlier trials) at a density of 200 ind. m”® (effective density of small shells in earlier trials, see the Results). Treat- ments (iii)-(v) were conducted using only U. tumidus, because both native species responded similarly in earlier trials (see the Results), and we wanted to limit the use of the legally protected and endangered A. cygnea. Furthermore, we tested the habitat preferences of S. woodiana for: (i) small bur- rowed conspecific shells (200 ind. m”) vs. pure sand and (ii) small burrowed con- specific shells vs. shells of native unionids (200 ind. m”) to check whether and how this species responds to shell beds. All the pairwise comparisons carried out within Experiment 1 are listed in Suppl. material 1: table S1. Experiment 2: Bivalve mobility and burrowing To test the effect of living S. woodiana and its empty shells on the locomotion and burrowing of A. cygnea and U. tumidus, we used tanks (40 x 30 x 35 cm) with a 10-cm layer of sand covered by the conditioned tap water (10 cm above the sub- stratum) (Suppl. material 1: fig. S2). As substrata, we used small S. woodiana shells (i) burrowed or (ii) lying on the sand surface, at a density of 200 ind. m”, as well as (iii) living S. woodiana (133 ind. m”) (Suppl. material 1: table $2). We used S. woodiana densities found to be avoided by the native unionids in Experiment 1 (see the Results). In the control treatment, bivalves were tested on (iv) pure sand without shells. A single substratum type was placed in each experimental tank. We introduced a single bivalve to the centre of the tank and recorded its behaviour using a CCTV camera (Samsung SNB-6004, South Korea) for 24 h. ‘The tests were replicated 15 times for each substratum and species. While watching the videos, we determined the following: (i) movement initiation time (time from the bivalve introduction to the first movement), (ii) locomotion duration, (iii) locomotion distance, (iv) locomotion speed (excluding periods of immobility), (v) duration of burrowing activity, (vi) mean burrowing level [%]. Every minute, we estimated the percentage of bivalve burrowing (using a 5-level scale: 0, 25, 50, 75 and 100%) by comparing the length of the part of the shell below the substratum surface with the part of the shell protruding above the substratum (according to Poznariska-Kakareko et al. 2021). Mean burrowing level was calculated according to the following formula: NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 248 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae 4 4 i=1 i=0 where: i — burrowing level: 5 steps ranging from 0 (totally exposed on the surface) to 4 (fully burrowed; t,— time spent by the mussel at burrowing level i. Statistical analysis Statistical analysis was carried out using SPSS 26.0 (IBM Inc.). We checked bivalve habitat preferences in Experiment 1 using y* tests of goodness of fit to compare their distribution within a given pair of habitats to a random distribution (as- suming equal numbers of individuals selecting each habitat). Because of the high departures of the mobility and burrowing data in Experiment 2 from normality and homoscedasticity assumptions (tested with Shapiro-Wilk and Levene tests, respectively), we compared bivalve behaviour (each species separately) on each substratum contaminated with S. woodiana to their behaviour on pure sand using non-parametric Mann-Whitney U tests with a sequential Bonferroni correction for multiple comparisons. Results Experiment 1: Habitat selection Both native species avoided small shells of S. woodiana (both burrowed and on the surface) at a density of 200 ind. m” (Fig. 1a, b, Table 2) and burrowed large shells at a density of 133 ind. m” (Fig. 1d, Table 2). Large shells on the surface were avoided at a density of 200 (A. cygnea) or 133 (U. tumidus) ind. m® (Fig. 1c, Table 2). Burrowed shells were avoided in favour of shells of the same size and density (200 ind. m* of small shells) located on the substratum surface (Fig. le, Table 2). Large burrowed shells were avoided by U. tumidus in favour of small burrowed shells of the same total volume (Fig. 1f, Table 2). The bivalves did not discriminate between living S. woodiana and pure sand (Fig. 1g, Table 2). Unio tumidus moved to the habitat formed by living S. wood- iana avoiding large shells burrowed in the substratum, whereas A. cygnea did not discriminate significantly between these habitats (Fig. 1g, Table 2). Unio tumidus showed a tendency to avoid shells of the native species, though it was non-significant (Fig. 2a, Table 2). Moreover, U. tumidus did not discriminate between shells of the native species and those of S. woodiana. Sinanodonta woodiana avoided conspecific shells and did not discriminate be- tween them and shells of the native unionids (Fig. 2b, Table 2). Experiment 2: Bivalve mobility Time from the introduction to the first movement of A. cygnea was not affected by the presence of shells and living individuals of S. woodiana (Fig. 3a, Table 3). Unio tumidus delayed the start of their activity in the presence of living S. woodiana (Fig. 3a, Table 3). Shells and living individuals of S. woodiana did not affect signifi- cantly the duration and distance of locomotion of A. cygnea (Fig. 3b, c, Table 3). NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 249 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae Habitat selection by Anodonta cygnea Habitat selection by Unio tumidus Small S. woodiana shells on surface f a | Pure sand] 133/m* Pure sand] 133/m’ (a) nd| 200/m’ 14] 200/m? (b) b Small S. woodiana shells burrowed Pure sand] 133/m* Pure sand| 133/m* | (c) 200/m* 200/m’ () Cc Large S. woodiana shells on surface Pure sand] 67/m* (e) Pure sand| 133/m’ 133/m’ (fF) 200/n @) d Large S. woodiana shells burrowed Pure sand] 67/m* Pure sand] 67/m* (h) 133/m’ 133/m Ci) e 200 small S. woodiana shells/m’* Burrowed Burrowed G/) if Large and small S. woodiana shells|of the same total volume | 133 large shdlis/m: (k) g Living S. woodiana 133 individuals/m* Pure sand] Living S. woodiana Pure sandjLiving S. woodiana C1) 133 large shells burrowed/m’| Living S. woodiana 133 large shells burrowed/m* (m) 60 40 20 | 0 20 40 60% 60 40 20 0 20 40 60% Percentages of mussels in particular habitats Figure 1. Habitat selection by A. cygnea and U. tumidus in the presence of substrata contaminated by S. woodiana in Experiment 1. Se- lected and avoided substrata are marked in green and red, respectively. The grey colour indicates non-significant differences. Blue letters in circles on the right refer to specific statistical tests presented in Table 2. a Habitat selection by Unio tumidus Pure sand] 200 small native shells /m? (n) 200 small native shells /m*| 200 small S. woodiana shells /m* (0) Habitat selection by|Sinanodonta woodiana 200 small S. woodiana shells /m* @) 200 small native shells /m*| 200 small S. woodiana shells /m* (0) 60 40 20 0) 20 40 60% Percentages of mussels in particular habitats Figure 2. Habitat selection of U. tumidus and S. woodiana in the presence of burrowed shells of native and invasive bivalves in Experiment 1. Selected and avoided substrata are marked in green and red, respectively. The grey colour indicates non-significant differences. Blue letters in circles on the right refer to specific statistical tests presented in Table 2. However, in the presence of all forms of S. woodiana contamination, distances travelled by A. cygnea were slightly shorter and movement duration longer, result- ing in a significantly slower crawling speed compared to that observed in the pure sand (Fig. 3d, Table 3). Unio tumidus increased duration and distance of their locomotion in the pres- ence of living S. woodiana or its shells on the surface (Fig. 3b, c, Table 3). In the pure sand and with burrowed shells, U. tumidus usually did not move horizontally at all, but burrowed immediately. Due to the total lack of locomotion of U. tumi- dus in the pure sand, it was not possible to calculate their speed on this substratum. NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 250 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae Table 2. Statistical analysis of habitat selection by A. cygnea, U. tumidus and S. woodiana in Experiment 1 (y? tests of goodness of fit comparing bivalve distribution within a given pair of habitats to the random distribution assuming no selection). Statistically significant differences are indicated by bold font and asterisks. y? — test statistic, P — statistical significance. —. k m n oO Anodonta cygnea | Unio tumidus | Sinanodonta woodiana Substrata control (pure sand) control (pure sand) control (pure sand) control (pure sand) control (pure sand) control (pure sand) control (pure sand) control (pure sand) control (pure sand) 200 small SW shells m? on surface xr a ¥ - xr P vs. | 133 small SW shells m? on surface | 2.13 0.144 | 2.13 | 0.144 - _ vs. | 200 small SW shells m? on surface | 8.53 | 0.003* | 4.80 0.028% - - vs. | 133 small burrowed SW shells m? | 6.53 0.068 | 4.80 | 0.273 - = vs. | 200 small burrowed SW shells m? | 10.80 | 0.001* | 6.53 | 0.011* 4.80 0.028* vs. | 67 large SW shells m® on surface - - 0.53 | 0.465 7 = vs. | 133 large SW shells m? on surface | 2.13 0.144 | 13.33 <0.001* - - vs. | 200 large SW shells m? on surface | 16.13 | <0.001* | — — — - vs. | 67 large burrowed SW shells m* 2B 0.144 | 1.20 | 0.273 = = vs. | 133 large burrowed SW shells m” 6.53 | 0.011* | 4.80 | 0.028* - - vs. | 200 small burrowed SW shells m? | 10.80 | 0.001* | 13.33. 0.000* - - 200 small burrowed SW shells m? | vs. | 133 large burrowed SW shells m” - — 6.53 | 0.011* — - control (pure sand) vs. 133 living SW m? Ze ES 0.144 | 0.53 | 0.465 = = 133 large burrowed SW shells m? vs. 133 living SW m? 3.33 0.068 | 4.80 | 0.028* — - control (pure sand) vs. | 200 small burrowed native shells m” - - 3.33 | 0.068 — — 200 small burrowed native shells m* | vs. | 200 small burrowed SW shells m? — _ 0.53 | 0.465 0.13 0.715 Experiment 2: Bivalve burrowing Anodonta cygnea spent a shorter time on burrowing in the substratum containing burrowed shells compared to the control sand (Fig. 4a, Table 3). There was no effect of S. woodiana habitats on the duration of burrowing activity of U. tumidus. The mean burrowing level of A. cygnea was reduced in the presence of burrowed shells and living S. woodiana (Fig. 4b, Suppl. material 1: fig. $2, Table 3). Unio tumidus responded to all types of S. woodiana habitats by reducing its burrowing level. Discussion In accordance with our first hypothesis, we reported avoidance of S. woodiana shells by native unionids. On the other hand, living individuals of the invasive species were not avoided even at a density twice as high (133 ind. m”) as the maximum densities observed so far in the field (Kraszewski and Zdanowski 2007). However, living S. woodiana did influence unionid behaviour: their presence de- layed initiation of activity and increased horizontal locomotion of U. tumidus, reduced locomotion speed in A. cygnea, and reduced burrowing of both species. An increased locomotion was also exhibited by U. delphinus Spengler, 1783 in the presence of the invasive clam Corbicula sp. (Ferreira-Rodriguez et al. 2018). It appears that the presence of shells induced displacement of native unionids, whereas living S. woodiana impaired the habitat quality for the natives, which tried to counteract by changing their activity. The increase in activity may be induced by searching for a habitat free of competitors, but its side effect may be the displace- ment of native bivalves to suboptimal environments, where they will be exposed to increased water flow or predatory pressure (Block et al. 2013). Another conse- quence of increased locomotion can be the depletion of energetic resources. As NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 251 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae Table 3. Statistical analysis of locomotion and burrowing of A. cygnea and U. tumidus in Experiment 2. Bivalve behaviour in the presence of S. woodiana shells (200 ind. m”, on the surface or burrowed) and living S. woodiana (133 ind. m~*) was compared to the behaviour of individ- uals exposed on the control pure sand with pairwise Mann-Whitney U tests. Statistically significant differences are marked with asterisks and those that are still significant with the sequential Bonferroni correction are marked in bold font. Z — test statistic, P — statistical significance. Variable gasses Anodonta cygnea Unio tumidus | Zz P Z P a | Movement initiation time | control (pure sand) | vs. | shells on surface | 2.30 0.022* 0.62 | 0.534 shells burrowed 152 0.129 Oat 0.443 | living individuals | 0.23 0.818 3.99 <0.001* b | Locomotion duration control (pure sand) | vs. slicliacontsurtace Ae 503 0.184 2.67 0.008* shells burrowed | LAG | 0.245 1.79 0.073 | | living individuals | 1.05 | 0.293, 2.40 0.017* c | Locomotion distance control (pure sand) | vs. | shells on surface | 1.00 0.319 | 2.67 0.008* shells burrowed aD, | 0223 179 0.073 i living individuals | _0.78 0.438 2.40 | 0.017% d | Locomotion speed control (pure sand) | vs. | shells on surface | 2.49 0.013* - - shells burrowed 2.44 0.015* - - | living individuals 2.10 0.035* - - e | Duration of burrowing activity | control (pure sand) | vs. shells on surface 0.73 0.467 1.81 | 0.071 | shells burrowed | 3.32 ~| ~— 0,001" 0.56 0.575 living individuals | 1.83 0.067 1.14 | 0.254 f | Mean burrowing level control (pure sand) | vs. shells on surface | 1.06 0.290 | 2.64 | 0.008* shells burrowed 3.11 0.002* 2.61 | 0.009* living individuals 2.63 | 0.009* 4.54 <0.001* 12007 ° a. Movement initiation time *K ___b. Locomotion duration 600 + a ° 10004 5004 ‘ 800 - o 4004 ° 2 600 4 | ° @ E = 300 + ; ; = = T Ata 200 4 2K ak *K 2004 c ° ao L — ° 0 = = 0 at sim =, a c. Locomotion distance , A d. Locomotion speed 504 1 J _s 064 4 = 25 3 307 5 0.4 o ° aa ° B 20 * e | ‘ : : a rf 2 0.24 3 10 | * * * ai = eS L [ « Bf I s. woodiane None Shells on Shells Living None Shells on Shells Living “ None Shellson Shells Living None Shells on Shells Living presence: ad surface burrowed mussels, - uae aioe mussels, a es bu TmOWRO mussels, CL guint pureed mussels, Anodonta cygnea Unio tumidus Anodonta cygnea Uhio tumidtis Figure 3. Mobility of A. cygnea and U. tumidus in Experiment 2: in pure sand (white bars), in the presence of S. woodiana shells (small shells, 200 ind. m”, blue bars) and in the presence of living S. woodiana (133 ind. m”, green bars) a movement initiation time b locomo- tion duration c locomotion distance and d locomotion speed. Asterisks indicate statistically significant differences in behaviour compared to that observed in the pure sand (see Table 3a-d for details of statistical test results). Boxplots present medians (horizontal lines), 1“ and NeoBiota 94: 243-259 (2024), DOI: 10.3897/neobiota.94.119622 252 Kamil Wisniewski et al.: The effect of shells and living Sinanodonta woodiana on native Unionidae a. Duration of burrowing activity | b. Mean burrowing level i 200 A *K lin 9 L 100 y | A : 2 = = SE a iol 2 1 Time (s) s) 5 3 2 Oo Qo ao o a Mean burrowing level (%) & S S * | | a —_§_x + ° 0 os S. weodiana None Shellson Shells Living None Shellson Shells Living None Shells on Shells Living None Shells on Shells Living presence: , surface burrowed mussels, surface burrowed mussels, re surface burrowed mussels, ' surface burrowed mussels, v ae 7 “7 ‘coi a aan —