The water quality of the water from Lake Constance used for the culture of the gammarids and the experiments remained equally good throughout the whole experimental period of 6 months (Table 1). The pH value did not change during the exposure time of 45 d in the experimental beakers, i.e. the PLA did not reduce pH by hydrolytic degradation.

In the 1^st run with new foils the mortality in the PET group was significantly lower than that in the PLA or control groups over the 45 days (RM ANOVA on Ranks, p: 0.027). After 46 days of exposure 40% - 55% mortality was reached in all treatments (Figure 1). In the 2^nd run with 45-day old foils this difference between the types of foils was supported, but not significant (p: 0.06). After 45 days of exposure 40% - 60% mortality was reached, with the lowest mortality in PET, followed by the control and PLA showing the highest mortality (60%) (Figure 1). In the 3^rd run with 90-day old foils, mortality in the controls was significantly lower than in PLA (p: 0.018), whereas in the 4^th run mortality was significantly lower in the control compared to both PET and PLA (p: 0.007).

At the end of the 1^st 45 d exposure, the weight loss of the alder leaves determined as dry weight loss did not significantly differ between the treatments (One-way ANOVA on Ranks, ns), (PET, PLA, control), whereby approximately up to 50% of the leaf was eaten (Table 2). At the end of the 2^nd experiment with 45 d old foils, the same results were found i.e. no significant difference in feeding on the leaves between the treatments. In addition, approximately 50% of the leaf was consumed by the gammarids within 45 days. At the end of the 3^rd experiment with 90 d old foils, up to 50% of leaves were eaten by the animals within 45 days of exposure. At the end of the 4^th experiment with 135 days old foils, the feeding rates were similar to the previous experiments (Table 2). The differences in

feeding activity between the treatments PET, PLA, and control were not significant in any of the 4 experiments (One way ANOVA: ns). Moreover, there was no significant correlation between the thickness of the leaves at the start and the observed weight loss at the end of the experiments, i.e. thinner leaves were not better food sources than thicker leaves, as originally hypothesized (Spearman Rank Correlation, ns). The degree of gut filling in the animals at the end of the experiments was always between 60% - 80% and showed no significant difference between the treatment groups (one Way ANOVA: ns).

The spontaneous locomotor activity of the gammarids was high during the whole experiment, both in the 1^st with new foils and the 2^nd with 45-day old foils, as well as in the 3^rd (90 d old foils) and in the 4^th (135 d old foils) runs (Figure 2). There was no significant difference between the treatments and the control in any of the 4 subsequent experiments (RM ANOVA on Ranks, ns.). Ventilation activity was significantly higher in PLA compared to other treatments (RM ANOVA on Ranks, p < 0.05) in the 1^st and the 4^th run (Figure 2).

In the 1^st run with the new foils, early colonization of the foils with biofilm within one week was observed, consisting of ciliates and flagellates. Diatoms and green algae, and esp. filamentous algae started to colonize the foils after more than 35 days (Figure 3). There were slightly more ciliates on PET foils (p: 0.07; paired t-test). For the other taxonomic groups, there were no differences between the type of foils. During the 1^st experiment, the colonization with diatoms was significantly lower than that with ciliates, flagellates and filamentous algae on both types of foils (one Way ANOVA on Ranks, PET: p: 0.008; PLA: p: 0.002). After 45 days some material degradation at the scratches for the numberings was observed on some PLA foils. The colonized biofilm did not differ in the composition before and after rinsing with tap water. However, the absolute numbers of the taxa were lower after rinsing, i.e. biofilm was partly loosely attached to both types of foils. Moreover, there was no weight loss of the foils (%; N = 12: PET: x −0.566, SD: 1.39 and PLA: x: 0.036, SD: 1.49), the high variation might be due to weighing errors.

In the 2^nd run on the 45-day aged foils the colonization scheme was similar to that of the 1^st run: ciliates and flagellates as early colonizers, followed by diatoms and green algae (Figure 3). Filamentous algae colonized after more than 30 days. Both types of foils were colonized equally with no statistical difference for any of the taxa. During the 2^nd experiment, the colonization with diatoms was significantly lower than that with ciliates and filamentous algae on both types of foils (one Way ANOVA on Ranks, PET: p: 0.002; PLA: p: 0.008). By the end of the

experiment, both foils were densely covered with filamentous algae as in the 1^st run. Rinsing with tap-water at the end of the 2^nd experiment did not change the biofilm taxa composition, but the absolute numbers were lower, i.e. the biofilm was loosely attached to both types of foils. The weight loss (%) of the foils at the end of the experiment was again very low and variable (N = 12: PET: Mean: 0.12; SD: 0.97 and PLA Mean: 0.84, SD: 0.81), Again, some further fissures at the numbering scratches in a few PLA foils could be seen in the 2^nd run.

Similar results were found in the 3^rd and 4^th run. Aggregation of organic matter associated with biofilm organisms increased during the 45 d of exposure, Filamentous green algae and flagellates were dominant on both types of foils and independent on the age of the foils (Figure 3). Throughout the 6 months of experiments, the biofilm colonization was reproducible in the 4 subsequent exposures of 45 d and did not differ between PLA and PET foils as substrate (paired t-test).

However, comparing one taxa group throughout the course of the 4 experiments on one type of foil, significant differences occurred for PET foils in the number of flagellates (p: 0.001), ciliates (p: 0.006) and filamentous green algae (p: 0.007). For PLA, foils significant differences occurred in the amount of flagellates (p: 0.001) and ciliates (p: 0.001) (RM ANOVA on ranks). This shows that there were seasonal differences in the quantitative composition of the biofilms throughout the 180 days in the water from Lake Constance used for the experiments (Figure 4).

The PLA foils especially showed damage due to handling, such as fissures at the scratches for numbering and deformation at the edges. This observation can be attributed to the fact that the PLA foil was a monolayer, hence degradation might start faster than in the PET GAG 3-layer foil. During the 4^th experiment, some PLA foils became intransparent, which might be due to hydrolysis. However, the visually observed damages were rare and of a minor extent, i.e. even the biopolymer PLA could not significantly be degraded within 180 days (at the end of exp. 4). The dry weight loss after the 4^th experiment showed slight increases such as 0.05% (SD: 0.02%) for PLA and 0.04% (SD: 0.06%) for PET, which might be due to weighing errors and/or biofilm colonization.

The potential toxicity generated by plastic foils (PLA, PET) was studied over 180 days in 4 subsequent experiments of 45 days. Mortality of G. fossarum exposed to new or aged (45, 90, 135 d) PLA foils was slightly higher compared to PET foils and controls. During the 4^th experiment with 135 d old foils, mortality was equally high in the treatments with PLA or PET foils, whereas control mortality was low.

Gammarids used the plastic foils as substrate and hiding places but did neither graze nor destroyed them. Toxicological studies on the effects of macroplastic on freshwater invertebrates are lacking, to our knowledge. Toxicity studies with microplastics are only of restricted relevance as microplastic toxicity is mostly caused by mechanical effects, such as clogging the gut system and tissue inflammation, where small microplastics pass the gut membranes [24] [25]. Weber et al. [26] studied the effects of PET microplastic (10 - 150 µm; 0.8 - 4000 g/ml) on Gammarus pulex and did not find any effects on survival after 48 days of exposure. Mateos-Cardenas et al. [27] showed that G. duebeni took up microplastic particles (1 - 1000 µm) adhered to Lemna minor during feeding, but to a lower extent (1 - 2 particles in the gammarid gut of 28.6% of the exposed gammarids). In a previous study, PLA particles (0.7 - 3 mm, grey, irregular shape) were not taken up by G. fossarum, probably due to the large size, compared to microplastic particles of PE, PP, and PBT, which were smaller in size (ca. 0.5 mm) [28]. During a 4 week’ exposure to these three different microplastic particles, no differences in mortality, feeding, and the locomotor activity of G. fossarum were found compared to the control group [28]. In the current study, spontaneous locomotor activity did not differ during any of the four subsequent experiments between the treatments, while ventilation activity was slightly higher in the PLA treatments. Increased ventilation is often regarded as toxic stress in aquatic animals, due to low pH or elevated metal levels [29]. This might indicate some migration of lactic acid out of the biopolymer; however, no changes in pH could be recorded. Feeding activity on the alder leaves did not differ between the treatments, which indicates no effects on the two types of plastic foils.

It appeared that especially in the beginning, during the 1^st experiment with new foils, PLA generated more potential for toxicity than PET. This might be due to the fact that PLA which is a monolayer biopolymer and might be more susceptible to chemical attacks as compared to the 3-layer copolymer PET-GAG. Potential migrants from PLA are monomers and dimers of lactic acid, which are regarded as environmentally safe as they represent common cellular components with rapid degradation [30]. Moreover, PLA foils lost transparency over 180 days of exposure and became opaque. This has been described as a change in the degree of crystallinity and was observed by Maiza et al. [31] in plasticized PLA experiments at 100˚C.

The toxicity of macroplastics might mainly arise from the release of toxic substances into the water. Evidence has been provided by studies on mineral water bottles, where some authors found traces of acetaldehyde and formaldehyde in water stored in PET bottles [32] [33]. Xeno-estrogens found in PET bottled mineral water showed effects in both the Yeast Estrogen Screen (YES hERa) test and in chronic exposures to the gastropod Potamopyrgus antipodarum (56 d). However, this might be due to contamination from the source of the water, as this finding could not be reproduced later [34]. PET caused cytotoxic and cytostatic effects on human lymphocytes, but no genotoxicity after 8 weeks of water storage [35]. Also, Ceretti et al. [36] did not find any genotoxicity (Microtoxtest, Micronucleustest, Comet assay) from migrating substances after 10 days in PET bottled and sparkling mineral water. PET-bottled water extracts (Solid Phase Extraction, SPE) did not induce either toxic or endocrine effects in in-vitro tests [37]. Migration of toxic compounds from plastics highly depends on the conditions and clearly increased with temperature, sunlight exposure and carbonation of the water [36] [37].

PLA showed faster degradation and was more prone to fissures than PET, also showing in-transparency and deformation of the edges within 180 days, whereas the 3-layer PET-GAG foil remained morphologically unchanged. PLA has been reported to adsorb water, followed by loss of transparency and increasing fragility [38]. PET is reported to be very stable for 15 years [39], with degradation times of 35 mm thick PET estimated to occur in 27 years [40] to 93 years [41]. The slow degradation of PLA during 180 days in our experiments might be due to low temperatures (10˚C) and neutral pH in the water from Lake Constance as well as the lack of sunlight during the exposures. Karamanlioglu et al. (2017) [15] reported a rapid microbial degradation of PLA in compost at 50˚C - 60˚C for 60 days, supported by findings of PLA-depolymerase activity at 60˚C in Bacillus strains [42].

Mutsuga et al. [43] found that pure PLA sheets lost about 0.28 - 15 µg/cm weight due to lactic migration after 180 d (at 40˚C). Weight loss of the foils was low in our experiments (<1%) and decreased from the 1^st to the 4^th experiment, i.e. with increasing age of the foils. The material loss might have been compensated by increased weight due to the adsorption of biofilm and water molecules.

The non-bacterial biofilm on PET and PLA foils consisted of early colonizers such as ciliates and flagellates followed by diatoms and filamentous green algae at last. Also, conglomerates of organic matter associated with biofilm organisms and extracellular polymers increased over time. There was no significant difference in the colonization of these taxa between the two types of glossy plastic foils.

Biofilm communities adhered to plastic polymers contribute to its degradation by microbial activities, such as degrading ester bonds via enzymatic hydrolysis. Microbes excrete metabolic enzymes, e.g. depolymerase and extracellular substances [44]. About 90 microbial genera have been found to degrade plastics, e.g. for PLA include Tritirachium album (proteinase K) which favors hydrolysis whereas a strain of Commamonas testosterone degrades terephthalate to protocatechuate, which is further degraded by Pseudomonas putida [44]. The composition of bacterial communities on plastics exposed to seawater differed between different polymers (esp. PVC compared to other plastics) and stages of biofilm succession, but these differences disappeared with an exposure time of 2 months [45]. However, Kirstein et al. [46] found only slight differences in biofilm on different types of synthetic polymers during 15 months’ exposure in seawater, with the exception of some polymer substrate-specific taxa.