Introduction
Leaf litter from forests supports complex detritus-based food webs in forest floors and streams through decomposition by the activity of microbial decomposers and invertebrate detritivores (Gessner et al., 2010). While the decomposition process has been mostly studied on monocultures, Seastedt (1984) suggested that litter mixtures might decompose at different rates from predicted from single species (i.e., showing nonadditive effects) due to the effect of resource heterogeneity on microbial decomposers and invertebrate detritivores. This catapulted research on the effect of mixing litter. Nonadditive effects may be due to leaf litter chemical characteristics, such as compounds which translocate from one leaf species and may enhance (e.g., nitrogen) or inhibit (e.g., polyphenols) decomposition of other species (Schimel & Hättenschwiler, 2007; Sanpera-Calbet et al., 2009; Lummer et al., 2012). Nonadditive effects may also be due to leaf litter physical characteristics, such as higher leaf toughness of one species which may act as an armouring effect against physical abrasion thus decreasing decomposition of other more fragile species (Wardle et al., 1997; Swan et al., 2008), and synergistic effects may also arise from the higher habitat/resource heterogeneity of mixtures when compared to single species (Epps et al., 2007), which enhances colonisation by biota and eventually results in higher overall processing rates by leaf consumers (Bastian et al., 2008).
The effect of mixing leaf species may reveal important consequences of tree biodiversity loss for the functioning and biodiversity of the recipient ecosystem (e.g., Gessner et al., 2010), but the subject is interesting per se , as similar studies, either in terrestrial, aquatic, or both ecosystems, often present strikingly contrasting results. This is well illustrated by five recent meta-analyses. Kou et al. (2020), using a data set from 65 studies in forest ecosystems across biomes concluded that N release (but not P) and mass loss were higher when litter was composed of two-three species. In contrast, Porre et al. (2020) found, in 78 studies across terrestrial ecosystems worldwide, that most litter mixtures had close to additive mass loss and only 15% of the data showed significant nonadditive mass loss in mixture, thus concluding that interactive effects among litter species are contextual and cannot be generalized beyond the context in which the results are obtained. Similarly, Chen & Chen (2021) found that the C:N:P ratios of plants, soil organic matter, microbial biomass and enzymes did not differ significantly between mixtures and monocultures in 169 studies in terrestrial ecosystems. Across biomes and ecosystems, Liu et al. (2020) compared 69 studies and concluded that the effect of mixing litter demonstrates nonadditive effects on decomposition rates in terrestrial and aquatic ecosystems, and that those effects are most frequently synergistic. Finally, Mori et al. (2020), using a data set from 151 studies across biomes and ecosystems found synergistic litter species-mixing effects for most ecosystems, including aquatic, but no effect in streams. Thus, the existence of a general, nonadditive relationship between litter mixing and the process of decomposition, both within and across terrestrial and aquatic ecosystems, remains elusive.
The flow of organic matter from the terrestrial to the aquatic habitats is the basis of energy and matter for many stream ecosystems (Wallace et al., 1997; Marks, 2019), with leaf litter falling vertically or entering the water by lateral movement through the forest floor (Benfield, 1997; Abelho, 2001). Abelho and Descals (2019) assessed leaf decomposition of alder, poplar and a 1:1 mixture of the two species in three scenarios across a gradient of terrestrial:aquatic exposures to determine if decomposition of leaf litter from lateral litter inputs to streams was similar to that of vertical inputs and concluded that decomposition rates, microbial and invertebrate colonisation differed significantly among exposure scenarios. Does the effect of mixing litter on decomposition differ between lateral and vertical inputs? Using the data set from Abelho & Descals (2019), the objective of this work was to assess (1) if the effect of mixing leaf litter on the process of decomposition (mass loss, associated decomposers and detritivores) differs between the stream ecosystem (under fully aquatic exposure) and when litter is exposed to a period of terrestrial exposure prior to immersion in the stream and (2) the global effect of the mixture across exposure scenarios. Based on the meta-analyses cited above, the tested hypotheses were: (a) the effect of the mixture is additive under fully aquatic exposure, given the results of Mori et al. (2020) for streams; (b) the effect of the mixture is synergistic for leaf litter with a period of terrestrial exposure, given the predominance of synergistic effects found by Liu et al. (2020) across ecosystems; and (c) synergistic effects are due to the enhancement of the decomposition process of poplar by the presence of alder, given the stimulating effects of N-fixing species on the decomposition of N-poor species (Hättenschwiler & Gasser, 2005).