Living in discarded containers: spiders explore a new niche created by littering in urban woodlands

Increased urbanisation is leading to littering of the environment. However, some animals may adapt to live in such altered habitats. The aim of this study was to assess whether discarded containers can serve as a suitable microhabitat for spiders. The study was conducted in 10 woodland areas in the city of Wrocław, Poland. In September 2018, a total of 939 containers were collected, of which 33.5% contained evidence of spiders having resided therein: webs, dead or living spiders, exuviae or cocoons. A total of 22 species and several other taxa that could not be recognised so accurately were detected. The most common of these belonged to Linyphiidae and Theridiidae. Juveniles (N = 103) dominated over adults (N = 58), and females (N = 34) were more numerous than males (N = 24). In 15 containers, interspecies pairs were found. Among ecological guilds, sheet web spiders dominated (60%) followed by space web spiders (24%), ground hunters (9%), ambush hunters (3%), specialists (3%), and other hunters (2%). Spiders were significantly more often found in colourless (38.3%) and green (35.5%) than in brown bottles (25.2%). They were also more numerous in glass bottles for sweet drinks (50.0%) than in beer bottles (28.0%), and slightly more frequent than in glass vodka bottles (33.3%). Our study showed that discarded containers are fully exploitable microhabitats for spiders and are used by these animals for at least three different purposes: hunting, hiding and breeding, however the effect on their fitness remains unknown.


Introduction
Increased human activity leads to drastic changes in the environment. Through habitat alterations, environmental pollution, and the spread or creating dispersal routes for invasive species and diseases, urbanisation is considered as a major factor of biodiversity loss (Maxwell et al. 2016;Sol et al. 2017). While the intensity and universality of these declines remains under debate, due to limited data, populations and species of many taxa are declining, sometimes dramatically (Ceballos et al. 2017;Didham et al. 2020;Montgomery et al. 2020). Although most animals cannot cope with human-altered environments, many others persist and have adapted to new conditions, including those in urban areas (McKinney 2008). Certain factors that occur only in urban areas can create favourable conditions for ecologically flexible species (Tuomainen and Candolin 2011). These include high spatial habitat heterogeneity (i.e. 'urban mosaic', Sattler et al. 2010), with some new microhabitats (Bar-Ziv et al. 2018), a warmer climate enhanced by the urban heat island (Youngsteadt et al. 2016), and artificial light at night (Mammola et al. 2018).
Spiders are characterised by high ecological plasticity and, as key predators in the ecosystem, play a crucial role in natural and urban habitats (Turnbull 1973;Riechert 1974;McIntyre 2000;Kralj-Fišer and Schneider 2012). Their reaction on urbanisation is highly species specific and depends on the individual nature of the city. Many assemblages of orb weavers and ground-dwelling spiders react to urbanisation by a decrease in species richness and an increase in abundance of selected species Lowe et al. 2016). There are, however, examples of spiders becoming less abundant and their communities less species rich as urbanisation increases (Dahirel et al. 2017;Piano et al. 2020). The success of some spider species in urban habitats may depend on their particular traits (Dahirel et al. 2017), their belonging to different ecological groups and inhabiting different habitat types (Piano et al. 2020), adaptation to new habitats such as building facades for spinning webs (Samu et al. 2004), or availability of prey Lowe et al. 2016). Spiders inhabit various urban microhabitats such as man-made structures (buildings, bridges, etc.;Heiling 1999;Mammola et al. 2018), cemeteries Kowarik et al. 2016), gardens (Burkman and Gardiner 2015), or parks (Lowe et al. 2016). It is also well known that some orb weavers and ground-dwelling species are attracted by artificial light, which is probably caused by high activity of invertebrates in the vicinity of lamps (Heiling 1999;Manfrin et al. 2017;Czaczkes et al. 2018;Mammola et al. 2018). Urban forests also provide a suitable habitat for many species. Indeed, Magura et al. (2010) and Horváth et al. (2012) found a higher diversity of ground dwelling spiders in urban forests in comparison to similar habitats in the suburban and rural areas, as a consequence of a high share of non-forest species therein. Thus, it seems that, in an urban matrix of various habitats, the number of xerophilous and light-preferring species penetrate from open habitats towards the more urbanised sites (Magura et al. 2010;Horváth et al. 2012;Möller et al. 2019).
Forests, especially those situated close to human settlements, are vulnerable to littering (Bartczak and Żylicz 2014). The commonly discarded items are beverage containers such as bottles and cans, which constitute a deadly trap for many small animals including mammals, reptiles, amphibians and invertebrates (e.g. Benedict and Billeter 2004;Poeta et al. 2015). Attracted by a smell of drink remains or decaying corpses, small-sized animals enter bottles or cans and cannot escape due to the slippery and wet walls. Kolenda et al. (2015) and Poeta et al. (2015) revealed that invertebrates were trapped in more than 40% of collected containers. Although, in the previous study, living arthropods were observed in the containers (Kolenda et al. 2015(Kolenda et al. , 2020, to the best of our knowledge, there are no studies concerning the use of this putative new niche for spiders to prey on trapped invertebrates. Only Prestwich (1977) observed complete webs, moulted exoskeletons and adult spiders inside discarded bottles.
In 2018, we started an ongoing project that is focused on testing the ecological trap hypothesis in relation to discarded containers. We supposed that some animal groups (e.g. ants; Kolenda et al. 2020) can dwell and breed inside these containers for at least part of their life, however the exact effect on their fitness remains unknown. Therefore, we assumed that spiders may also use discarded containers for different purposes, such as foraging or reproducing. We hypothesised that the trapping of potential prey in such places may attract spiders to spin their webs inside, for capturing the lured insects, similarly to web building near artificial light source. Consequently, we assumed that not the active or ambush hunters, but the web-building species are the most common predators in this artificial microhabitat. Among them, females and juveniles should dominate, while males usually actively explore the area in order to find a mate (Burgess 1978). Alternatively, spiders could use the discarded containers simply as one of many shelters available in the forest litter, without any special preferences towards them. Thus, our research is the first step in exploring discarded containers as potential ecological traps for spiders. It will help in planning further research that assesses how these litter items influence spiders' fitness, data crucial for distinguishing the ecological traps (Hale and Swearer 2016).

Material collection
Spiders dwelling in discarded containers were studied in Wrocław (Poland). Ten forest patches were selected within the boundaries of the city. Total area of each site varies from 16 to 139 ha. The forests are located mainly in river valleys. Most of them are formed by oak-hornbeam and riparian habitats; however, some were strongly altered and remain under high human pressure. In September 2018, we collected uncapped, discarded containers along forest paths within a 10-m-wide transect starting from the entrance to the forest. If possible, we picked up around 100 containers per forest. In total 939 bottles or cans were sampled. The containers were secured against spillage and transported to the laboratory. Afterwards, the presence of spider webs was noted and the content of containers was analysed on a plastic tray in order to find living and dead spiders, as well as their exuviae and cocoons. These were identified to the lowest possible taxonomic level, and all taxa that were assigned at least to the family level were grouped according to ecological guilds, according to Cardoso et al. (2011). The only exception were three individuals of Pachygnatha listeri; although classified as orb-web, they abandon their webs as adults and migrate (Martin, 1978), exploring their environment in search of mating opportunities. Thus, we removed them from this analysis. The basis for species identification were keys by Nentwig et al. (2020) and Roberts (1987). Spider names are in accordance with the World Spider Catalog (2021).
As the overall number of dead spiders was low (see "Results" section) all individuals (alive and dead) were pooled in one dataset. Due to (i) the exploratory nature of the study, and (ii) a lack of knowledge about the effects of container features on the invertebrate occurrence and mortality, we intended to take into account a variety of their traits: capacity (CAP), opening diameter (DIA), material (MAT), presence of a neck (NECK), colour 1 3 (COL), and original content (CON) (for details see Online Resource Table S1). These features can be considered as important for invertebrates as they seem to affect attractiveness (colour, material, size, content) as well as accessibility of the containers (size, presence on neck). However, due to (i) small sample size of some classes and (ii) overlapping of some variables (e.g. COL nested in MAT in the case of cans), for statistical analyses we used only selected variables, values of categorical variables, and variable sets (see Online Resource Table S1, S2 for details). The main criterion for including the data for the analyses was the material used in container production ("Plastic" or "Glass" or "Aluminium"). An additional criterion for COL and CON traits was the size of the subclass larger than 50. In the case of CON, the subclasses were unevenly represented between material types, and in some cases, the subclass size amounted to zero. For example, "vodka" (including other strong alcoholic drinks such as whiskey, bourbon, etc.) was not noted in plastic, while "water" was not noted in glass containers.

Statistical analyses
Chi-squared tests were used to: (i) compare the frequency of adult and juvenile individuals as well as females and males; (ii) test the differences in the frequency of spiders presence (webs, dead or living spiders, exuviae or cocoons) between the container classes as well as the differences in composition of the most numerous ecological guilds between the container classes. In cases where subclass size was smaller than five, Fisher's exact test was used. These analyses were carried out using Statistica 12 software (StatSoft Inc. 2014).
The dependency of spider presence and the dependency of spider guild representation on the container features were analysed with the use of multiple generalised linear models with binomial error frequency distribution, as well as logit link function. Due to the above mentioned limitation, linked with the small size of some subclasses, only a single categorical variable (each of four variables: MAT, NECK, COL, CON) and two numerical variables (CAP and DIA), together with the interactions between the categorical variable and each of two numerical variables were considered when building the models. For example, in the case of the material (MAT), five predicator variables were used for model building: MAT, DIA, CAP, MAT × DIA and MAT × CAP. We assumed the high homogeneity of the studied sites as the habitats were similar. Although spider species composition sometimes differs in particular sites even where the habitats are the same, we considered this heterogeneity of data should be much less pronounced as only the presence or absence of spiders, and the ecological groups are analysed. The best models were selected with the use of R package MuMIn (Bartoń 2016) by the corrected Akaike Information Criterion (AICc) (Burnham and Anderson 2004). When a subset of the best models consisted of more than one model matching the criterion ΔAICc < 2 (the difference between the AICc of the best model and the AICc for each of the other models), the models were averaged with the use of package MuMIn. The significance of a model's coefficients was assessed with the aid of estimated statistical significance (P) and 95% confidence intervals (CI). In the case of averaged models, the number of best models (N models) containing a given predictor was defined, as well as the relative variable importance (RVI) of the predictors being estimated. The analyses described in the paragraph were performed in the R environment ver. 3.5.2 (R Core Team 2018).

Spider composition found in containers
We found evidence of spiders presence-webs, dead or living spiders, exuviae or cocoons-in 33.5% (N = 315) of all collected containers (N = 939) (Fig. 1). Webs were the most common proof of the occurrence of spiders in bottles or cans; they were found in 30.9% of samples (N = 290). However, only 40.7% of these (N = 118) also contained spiders. In general, a total of 182 spiders have been found in 15.5% of containers (N = 146), including 163 alive and 19 dead individuals (Online Resource Table S3). Additionally, a total of 26 exuviae and 2 cocoons were found in 22 and 2 containers, respectively (Online Resource Table S3). In two cases, the exuviae and spider from the same container belonged to the same species, thus we assumed that they represent one individual, which was taken into account in further analysis.
Except for 6 spiders and 9 exuviae, the other specimens or traces of spiders were classified at least to the family level. The most common were Linyphiidae and Theridiidae, which were represented by 113 and 43 individuals, respectively (including living or dead spiders, Fig. 1 Examples of containers utilised by spiders: a a glass bottle with Neriene clathrata (Linyphiidae) and its web, b a plastic bottle containing prey attached to the web 1 3 exuviae, and cocoons) (Fig. 2a). Among Linyphiidae, Tenuiphantes spp. dominated, while Enoplognatha spp. were the most numerous theridiid spiders (Online Resource Table S3). Remaining individuals were classified to 11 different families, which were represented by  Table S3 for a detailed list of spiders, their exuviae and cocoons).
Most containers contained only one individual; however, in four cases, we found two or more spiders: (i) Tenuiphantes spp. female and juveniles, (ii) T. flavipes male and juveniles, (iii-iv) T. flavipes female and a juvenile. Additionally, there were 15 interspecies pairs found in a single container (Online Resource Table S4).
In multivariable analysis for the presence of spiders in relation to container features, among the numeric features only CAP was present in some of the best models (5/9), categorical variables were present in all but one model, and interactions between numeric and categorical variables were absent with one exception (MAT × CAP) (Online Resource Table S6). The categorical effects were more significant (in terms of RVI, P, and N models) in the averaged models (Table 2). Spider presence in containers was positively (with marginal statistical significance) affected by NECK, by green or colourless glass (when Table 2 Averaged models of relationships between the presence of spiders in containers with six predictors (container features): four categorical predictors-material (MAT), presence of neck (NECK), colour (COL) and contents (CON), and two numeric predictors-capacity (CAP) and opening diameter ( 1 3 compared to brown) and by sweet contents (when compared to beer or vodka, but with marginal statistical significance). These effects, however, were present in the models with very poor Goodness-Of-Fit as reflected by low values of Nagelkerke-R 2 , amounting to 0.000-0.026. In the analysis of the effect of container features on the representation of spider guilds (sheet vs. space web spiders were only analyzed), one numeric variable (CAP) was also present in the best models (4/7) besides categorical variables (5/7) (Online Resource  Table S7). In the averaged models (Table 3), the presence of sheet web spiders was affected negatively by plastic containers, however at a positive interaction MAT (plastic) × CAP. In the case of other categorical variables, most effects in the averaged models were insignificant with exception of the negative interaction NECK × CAP. Also, Goodness-of-Fit of these models was poor. The Nagelkerke-R 2 amounted to 0.000-0.130 in individual best models (Table S7), which indicates a weak predictive ability of these models. Table 3 Averaged models of relationships between the presence of web spider guild (space vs. sheet web) in containers with six predictors (container features): four categorical predictors-material (MAT), presence of a neck (NECK), colour (COL) and contents (CON), and the two numeric predictors-capacity (CAP) and opening diameter (

Ecological significance of discarded containers
This study revealed that discarded containers are frequently dwelled by various groups of spiders. Detailed discussion with previous results is hardly possible because there is almost no research that shows influence of discarded containers on arthropod populations or communities. However, in contrast to our recent findings on ants (Kolenda et al. 2020), the containers serve for spiders more as an artificial microhabitat than as a deadly trap. Indeed, the number of containers with dead specimens was very low (1.7% of all), especially in contrast to ants (10.3% of containers with dead workers, and only 4.4% used as a nest; for details see Kolenda et al. 2020). Moreover, Lavers et al. (2020) found that discarded plastic bottles are a lethal trap for crabs on beaches, while Kolenda et al. (2015) and Poeta et al. (2015) noted mostly dead beetles and molluscs inside containers collected from suburban forests and sandy coastal dunes, respectively. Discarded containers can, however, act as suitable habitat for animals in aquatic conditions. In a lowland dam reservoir, macroinvertebrates diversity found in the bottles was as high as on phytolittoral bottom and higher than on other natural or artificial studied substrates (Czarnecka et al. 2009). Probably such items provided shelter against harsh environmental conditions of the near-shore zone of the reservoir (Czarnecka et al. 2009). Some spider traits may explain why they did not die so often in the garbage traps but rather inhabited them, even if the containers were not dry. Many spiders readily walk on steep surfaces due to some adhesive structures (the so-called scopulae in wandering spiders) and support themselves with security threads in case they fall down. Many of them walk quite easily on the surface of water using surface tension; they may, in fact, use numerous techniques to move on liquid surfaces (Stratton et al. 2004). Similarly, some Linyphiid species easily avoid being trapped and readily walk on steep walls of pitfall traps (Topping 1993) and even build webs within (so exploring artificial habitats similar to the containers we studied). As predators, the spiders seem also not to be lured into the traps by the water and decaying matter, as some beetles or small mammals are (Benedict and Billeter 2004;Kolenda et al. 2015).
We observed that spiders which dwell in discarded containers use them for at least three different purposes: hunting, hiding, and breeding. As a confirmation of the latter, we found adult couples and cocoons. The only cocoons we found belonged to Ero spp., which are commonly found in European forests where they are attached to plants (Finch 2005a), e.g. to tree trunks. Their presence in containers can be additionally explained by the fact that Ero are small predators of the web spiders, using aggressive mimicry for hunting spiders sitting in webs (Czajka 1963). Discarded containers are also a presumably convenient moulting place for several spider families, even those that we did not record from live specimens; for instance, Gnaphosidae. This shows that containers serve as hiding places for some wandering spiders. However, this role cannot be overestimated, as spiders can use any shelter places in the forest litter. They may also serve as a hunting site for some spiders. Many studies have shown that prey availability is often crucial for spiders locating their webs on a site (e.g. Samu et al. 1996;Harwood et al. 2003; see also Wise 1993). Nearly 1/3 of collected containers contained webs. In many randomly checked containers in the field, we found prey remains in these webs (Fig. 1b).
Despite the observation of various life stages and low mortality of spiders, as well as the fact that the containers are used by spiders for different purposes, we cannot conclude that they are a suitable habitat for these arthropods. Further studies comparing reproductive and hunting success, or species richness and specimen number in the containers and in their surroundings are necessary. This could explain which species of spiders avoid containers and which of them take any advantage of inhabiting the containers or fall into ecological trap.
The presence of more than one web in a single container does not necessarily mean that each of them was built by a single spider. As the example of Tenuiphantes tenuis shows, one specimen may build several webs and one web may be used by several specimens consecutively (Samu et al. 1996). The web spinners are often characterised by a kind of 'floating populations' (as shown by a sheet-web spider Linyphia triangularis; Toft 1998), due to frequent web takeover and the fact that empty webs are often colonised. We assume that this could be the case with some of the observed species. We cannot also exclude the occurrence of competition between spiders within the containers. In several cases, we have observed inter-and intraspecies pairs of spiders inhabiting one container. Some studies reported a high level of competition for web sites in the field (Samu et al. 1996;Heiling and Herberstein 1999;Riechert and Hall 2000;Hardwood et al. 2003). Related linyphiid spiders often have very similar niche preferences (e.g. a preference for web positioning; Toft 1987). Some observations showed that it is a common phenomenon that a heavier or larger spider overtakes the web (Samu et al. 1996;Eichenberger et al. 2009) and intraguild predation is also present among spiders (Finke and Denno 2002). Small linyphiid spiders may not be deterred from entering the site even in the presence of chemical clues from the larger spider predator (Wetter et al. 2012). However, some data shows that smaller spiders avoid places with the scent (the kairomons) of a larger spider predator (Persons and Rypstra 2001).
The open question is whether the containers provide conditions suitable for overwintering. Generally, in temperate zones, arthropods are prone to mortality due to low temperature; thus, they overwinter in habitats with stable conditions (Roume et al. 2011). Some, however, present adaptations to survive and are active even at temperatures below freezing-especially when prey is available (Aitchison 1984;Korenko et al. 2010;Lee 2012). Forests are important overwintering habitats for arthropods. Some species, including those found in our study (e.g. belonging to genera Enoploghatha, Neriene), overwinter in litter (Martyniuk and Wise 1985;Nähring 1991). Furthermore, whole season observations are required to assess if containers are left by spiders before winter or pose a year-round habitat.

Fauna overview-species filtering and biases
We recorded an array of typical litter-and undergrowth-dwelling species of the forest litter of the European temperate zone (Stańska et al. 2002;Finch 2005b;Milasowszky et al. 2015;Košulič et al. 2016), but only three of them dominated. The most dominant spider of these "garbage-assemblages" was Tenuiphantes flavipes, a very common species from Linyphiidae, which predominantly lives on the ground surface in different woodlands (Hänggi et al. 1995) where it builds small sheet webs. The species has a slight preference for shade and a stronger one for moist habitats (Entling et al. 2007) but its niche preferences and the habitats it utilises are broad (Hänggi et al. 1995;Entling et al. 2007). We observed the dominance both of the juveniles (not identified with certainty, but many of them most probably belonged to this species) and of adult specimens. This species builds webs for hunting and mating; we, therefore, suppose that it explores the garbage for both purposes.
The second most common spider was Enoplognatha spp., most probably E. ovata, which is confirmed by habitat type (Oxford 1992;Barthel 1997) and the presence of an adult specimen. This spider builds webs in the understory of different habitats, including forests, and prefers dense vegetation. This refers, however, to the adult specimens and not necessarily to the juveniles-spiders may shift microhabitats with successive life stages. The juveniles, which predominated in our samples, are known to overwinter in litter and disperse in spring into higher strata of the understory (Nähring 1991). A similar rule most likely pertains to other relatively big species of Linyphiidae, such as Neriene clathrata or Linyphia hortensis, which are common in our samples. For instance, Neriene clathrata generally weaves its sheet web in the lower strata of forest undergrowth (Wright and Coyle 2000).
When considering guild composition, there were hardly any ground or ambush hunters in the containers; however, we noticed some signs of their presence (the exuviae) and a few dead specimens belonging to these ecological groups. The wolf spiders (Lycosidae) are known to have a 'sit-and-move' strategy of exploring prey in their environment (Samu et al. 2003). Therefore, we suppose that they may be only the temporary visitors of containers, performing in this case a 'come-and-go' strategy. The predominance of sheet and space web weavers in our samples may result from the fact that these groups may easily adapt to structural properties of habitats and spend more time in discarded containers while waiting for their prey or while mating. Some of them even enlarge their webs with time by overbuilding the older parts of their snare (Benjamin and Zschokke 2004).
Without comparative study of species from the surroundings, we cannot precisely determine if the dominance of spider species in the containers results from their high abundance in the environment or from some species filtering towards the new microhabitat. Moreover, a total of 22 species and several further taxa that could not be recognised so accurately is a low number for typical temperate forests (Scharff et al. 2003;Stańska et al. 2016), which may reach about 80 for a sampling plot when sampled extensively (Scharff et al. 2003). However, the intensity of sampling and the used method was different in the studies, so our data are hardly comparable. The environmental factors, which influence richness and composition of forest spider assemblages, are the canopy cover, humidity, litter structure, and plant structure and density (Bultman and Uetz 1982;Samu et al. 1996;Pearce et al. 2004;Entling et al. 2007;Oxbrough et al. 2010); however, they were not considered in this study.
It should be also noted, that other factors can affect species richness. For instance, each sampling method used in arthropod studies is somehow biased (Topping and Sunderland 1992;Prasifka et al. 2007). Typical for the pitfall traps is a high dominance of spiders belonging to Lycosidae, the large ground hunters, which-along with some other families of wandering spiders-are very active and, as a result, dominate the yield of pitfall traps (Uetz and Unzicker 1976). On the contrary, this family was hardly present in our samples and recorded from dead specimens (or exuviae). The sampling period is also important. Spiders show a strong pattern in their phenology (e.g. Riecken 1999;Hsieh and Linsenmair 2012;Blandenier et al. 2013). Our study was conducted in September and the highest species richness is recorded in spring and early summer (Niemelä et al. 1994).
The other factor which we did not consider is spiders' circadian activity (see Krumpálová and Tuf 2013 for data from central European forests). Some hunters which are active during the day time may have been absent in our samples as we sampled the containers during those hours. Thus, we do not claim that our list covers the whole diversity of spiders being entrapped or utilising the litter left over in the forests.

Preferences toward containers
Spiders are the generalist terrestrial predators which explore every possible space niche that is suitable for them to forage or breed. Our results indicate that there was no clear preference of spiders for any type of discarded containers. One of the exceptions was the low preference towards brown glass bottles and those containing beer or high-volume alcohol. This result should be treated with caution, as it is very hard to relate it to spider biology and the sample size was small in some subclasses. More data and further (meta)analysis could probably show clear preferences of spiders towards container types. We assume that there might be some preferences, based on studies analyzing efficiency of pitfall traps that are used in ecological studies. Methodological works on the construction of pitfall traps suggest that spiders catching rate may vary depending on their features. Namely, the effectiveness of these traps might be affected by the presence or absence of covers (Buchholz and Hanning 2009), the colour of traps (Buchholz et al. 2010), the entrance size (Work et al. 2002) or properties of fluid inside the container (Pekár 2002;Schmidt et al. 2006). Therefore, we had anticipated some bias in the number of spiders according to specific container categories.
The other factor that might have influenced spiders could be the shape of the containers, which we did not analyse, because web-building spiders are dependent on the availability of sites for spinning their snare. On the other hand, the linyphiids may also easily adapt the shape of their webs to the available spatial conditions (Samu et al. 1996;Rybak 2007). The position of the container opening could also be significant, because forest litter is not a single layer-it is a diversified three-dimensional space which creates several microhabitats for spiders (Wagner et al. 2003) and some spiders may even migrate into very deep soil layers (Laška et al. 2011). However, the lack of any explicit influence of container features on spiders, suggests that these animals-in contrast to many other arthropod groups-may not be attracted to them and simply exploit every available microhabitat of the forest litter.

Conclusions
The goal of our research was to present another example of wildlife adapting to life in a human-modified environment. Beverage containers are a common litter items in land and aquatic environments around the world (Roman et al. 2020;Kolenda et al. 2021). They may become fully exploitable (micro)habitats for some spiders. However, further studies, i.e. concerning presumptive overwintering or hunting and reproductive success, should explain if containers are really a favourable habitat for them or constitute an ecological trap for spiders.
Despite potential positive effect on some selected spider and ant populations, it should be noted that discarded containers do not promote rare species and act as lethal trap for other arthropod groups and small vertebrates (Benedict and Billeter 2004;Kolenda et al. 2015Kolenda et al. , 2020Poeta et al. 2015;Lavers et al. 2020). As our recent study showed, bottles and cans are also a threat for medium size animals such as deer, red foxes or weasels (Kolenda et al. 2021). Moreover, such microhabitats favour the spread of the yellow fever mosquito, Aedes aegypti-a dengue vector (Juarez et al. 2020). Therefore, as a significant threat to ecosystems, they should be regularly removed from the environment.