Leaf litter quality drives the feeding by invertebrate shredders in tropical streams

Abstract Amazon and Cerrado‐forested streams show natural fluctuations in leaf litter quantity along the time and space, suggesting a change on litter quality input. These natural fluctuations of leaf litter have repercussion on the organic matter cycling and consequently effects on leaf decomposition in forested streams. The effects of the quantity of leaf litter with contrasting traits on consumption by larvae of shredder insects from biomes with different organic matter dynamics have still been an understudied question. The Trichoptera Phylloicus spp. is a typical shredder in tropical headwater streams and keep an important role in leaf litter decomposition. Here, we assessed the consumption by shredder Phylloicus spp., from Amazonia and Cerrado biomes, on higher (Maprounea guianensis) and lower quality leaves (Inga laurina) in different proportions and quantities. Experiments were performed concomitantly in microcosms approaches, simulating Cerrado and Amazonian streams. Higher leaf consumption occurred in Cerrado microcosms. Litter quantity influenced negatively leaf consumption by shredders in Cerrado, in opposition to Amazonia, where consumption was not affected by leaf quantity. In both sites, we observed higher consumption by shredders in treatment with only M. guianensis and no difference between other treatments with mixture of leaves. In treatment with litter of I. laurina, we noted the use of substrate for case building (due to the higher leaf toughness), affecting the fragmentation process. Therefore, our results indicate that leaf litter quality drives the preference of consumption by Phylloicus larvae in Cerrado and Amazonia streams.

precipitation as a limiting factor (Tonin et al., 2017). These natural seasonal changes on litter input suggest a change of litter quality input (Gonçalves & Callisto, 2013;Tonin et al., 2017), as well as repercussions for the flow of energy and the cycling of organic matter in streams (Bueno et al., 2016).
Riparian zone closes the canopy above streams (Gonçalves & Callisto, 2013), reducing the light input in lotic ecosystems and, consequently, the primary productivity of the system (Vannote, Minshall, Cummins, Sedell, & Cushing, 1980). Therefore, the main source of energy for low-order lotic systems comes from the input and subsequent decomposition of litter material originated from the riparian plants, between 60% and 80% Esteves & Gonçalves, 2011) derive from vertical and lateral contribution of the stream edges, respectively . Leaf litter in these ecosystems can be retained and accumulated on the riverbed , where can undergo leaching of soluble compounds (e.g., tannins and polyphenols; Graça et al., 2015) and be colonized by microorganisms in the conditioning process (Graça, Hyde, & Chauvet, 2016). Finally, invertebrates fragment leaves to use them as food resource and/or shelter (Gessner, Chauvet, & Dobson, 1999;Graça et al., 2015). Detritivore invertebrates (mainly shredders) can dramatically increase leaf breakdown rates Moulton, Magalhaes-Fraga, Brito, & Barbosa, 2010) showing the importance of the shredders for functioning of headwater streams systems (Graça, 2001).
In general, shredders have been found in low abundance in tropical streams (more adapted to lower temperatures; Boyero et al., 2011;Prather, 2003), but due to high biomass and body size they can be important for cycling of organic matter and energy flow in aquatic ecosystems (Martins, Melo, Gonçalves, & Hamada, 2015;Rezende, Petrucio, & Gonçalves, 2014;Tonin, Hepp, Restello, & Gonçalves, 2014). The genus Phylloicus Müller, 1880 (Trichoptera: Calamoceratidae) has approximately 60 species, distributed from South to Central America. Phylloicus' larvae are commonly found on submerged leaf litter (Prather, 2003). It is the most abundant shredder taxon in Cerrado (Rezende et al., 2014 and Amazonian headwater streams (Martins, Melo, Gonçalves, & Hamada, 2014;Martins et al., 2015). Phylloicus is a typical shredder and uses leaf litter deposited in pool areas on streams bed to obtain their food and material to case construction (Wantzen & Wagner, 2006). In this process, Phylloicus larvae convert the coarse particulate organic matter into fine particles and dissolved organic matter (Prather, 2003). This highlights the importance of understanding factors that change the leaf litter processing by Phylloicus, and consequently, the functioning of ecosystem. Ferreira et al. (2015) analyzed the diet of Phylloicus larvae and pointed out that the particulate organic matter (fine and coarse) consumed varied among instars with the proportion of food related to stream characteristics. These results indicated the complex diet and the direct connection with local and temporal resource availability (input of leaf litter), highlighting the need for further studies. Events such as temporal changes of litter quantity and the physical environment have been claimed to affect Phylloicus populations (Leite, Silva, Navarro, Rezende, & Gonçalves Júnior, 2016;Rezende, Santos, Medeiros, & Gonçalves, 2017). As well, intrinsic leaf traits, such as leaf litter size (Rezende et al., 2018), hardness and nutrients (Biasi, Cogo, Hepp, & Santos, 2019;Graça, 2001), lignin (Gessner, 2005), and phenolic concentrations (Moretti, Loyola, Becker, & Callisto, 2009) are factors that denote the leaf litter quality and explain shredders feed preference. However, biological interactions as competition  and risk of predation (Navarro, Rezende, & Gonçalves, 2013) can affect negatively the population dynamics of Phylloicus.
Leaf litter processing by Phylloicus larvae can also be negativity influenced by environmental conditions, as increase of temperature (Martins, Melo, Gonçalves, Campos, & Hamada, 2017;Martins, Rezende, et al., 2017;Navarro et al., 2013). Nevertheless, all these mentioned studies just refer to single populations, using leaf litter and larvae just from the same place in their design.
Previous studies just investigated food preference of Phylloicus larvae (Moretti et al., 2009;Navarro et al., 2013). However, the effects of leaf litter quantity of different leaf traits on leaf litter consumption in biomes with different organic matter dynamics have still been an understudied question. Our aim was to assess consumption by Phylloicus larvae from different geographic origins on higher (Maprounea guianensis) and lower quality leaves (Inga laurina) of different proportions and quantities under controlled conditions of microcosms. This experiment will allow assessing of possible impacts on leaf litter processing and survival of Phylloicus spp. by changes in Amazonia and Cerrado riparian forest composition. Our hypotheses were: (a) Due to the shredder feeding preference for higher quality leaf litter (Biasi et al., 2019;Rincón & Martínez, 2006), a superior consumption will be recorded in more M. guianensis proportion treatments; (b) A higher leaf litter availability will decrease consumption of harder leaf litter, because consumers tends to potentialize the feed in resources that are energetically better for metabolism maintenance (Kaspari, Donoso, Lucas, Zumbusch, & Kay, 2012).

| Phylloicus' sampling and identification
Phylloicus spp. larvae ( Figure 1) were sampled by "active searching" (saw target species and then collected by kicknet) by 4 hr of sampling effort (single day). For identification, some Phylloicus larvae from Cerrado stream were collected in the field and kept, individually, in containers with small portion of leaves collected in the field and stream water with constant aeration, each container was covered with fine mesh of 0.5 mm held by a rubber band. Daily observations were made until newly hatched adults were observed.
They were captured, preserved in 80% alcohol, and sent to the National Institute for Research in Amazonia (INPA) for identification. The specimens from Cerrado represent an undescribed Phylloicus species. Larvae collected in the Amazonian stream belong to the species Phylloicus elektoros Prather, association with adult was made by rearing procedures, however, the larva is not formally described yet.

| Laboratory procedures
The experiments were performed in the microcosms of the Laboratory of Limnology/AquaRiparia in Department of Ecology at the University of Brasilia (UnB) for the Cerrado biome experiment, and in the Laboratoty of Citotaxonomia e Inetos Aquáticos located at the Instituto Nacional de Pesquisas da Amazônia (INPA) for Amazonian biome experiment. Phylloicus larvae were taken to the laboratory in cool boxes and placed in containers (15.5 cm × 15.5 cm × 12 cm, 2 L volume) with mineral water and previously calcined gravel (in an oven for 4 hr at 550°C) on the bottom (Martins, Melo, et al., 2017;Martins, Rezende, et al., 2017). Larvae were kept in containers in an experimental room with continuous aeration at a temperature of 20°C and with a light/dark ratio of 12/12 hr throughout the experiment. Daily, we measured dissolved oxygen (Lenway 970 Meter DO2; Cerrado: 6.9 ± 0.3 standard error mg/L; Amazonia: 6.6 ± 0.22 mg/L), electrical conductivity (Lenway 430 pH/cond. Meter; Cerrado: 20.1 ± 1.5 µS/ cm 2 ; Amazonia: 72.8 ± 5.4 µS/cm 2 ) and pH (Lenway 430 pH/cond. Meter; Cerrado: 6.5 ± 0.04; Amazonia: 4.7 ± 0.1) in all containers.
Phylloicus larvae had their cases removed before the start of the experiments to avoid the consumption of external organic matter (e.g., consumption of case).
Leaf litter of Maprounea guianensis and Inga laurina were conditioned in litterbags (0.5 mm mesh) for seven days at Capetinga and Barro Branco streams for Cerrado and Amazonia experiments, respectively. As described for Gomes, Medeiros, and Gonçalves (2016), Maprounea guianensis (higher quality leaf litter) presents 23% of lignin, 15% of cellulose, 36% of polyphenols, 0.76% of Nitrogen Subsequently, the conditioned leaves were cut into disks (1.98 cm diameter) and freeze-dried (Terroni, LT-AISI 304 model). Sets of disks were weighed on a precision balance (0.01 mg; Shimadzu, AUW220D model) to determine the initial dry mass and distributed among the treatments. At the end of the experiments, the leaf disks were removed, dried at 60°C for 48 hr and subsequently weighed to obtain the final dry weight. The mass leaf loss (MLL) between treatments was calculated by the difference between the initial and final dry mass, after seven days of experiment in both locals (Figure 1).

| Experimental design
In the experiment, we used 80 containers (Cerrado

| Statistical analysis
Through Factorial-ANOVA between biomes resource availability (leaf quantity), litter proportions (leaf quality) and interaction of these factors (independent variables), we tested the percentage of leaf litter used as food resource by Phylloicus (dependent variable). The data were transformed whenever necessary with arcsine square root of the ratio to obtain the best fit (Crawley, 2007). We used contrast analysis to discriminate among statistically significant categorical variables (Rezende et al., 2018). All analyses were conducted in R (R Core Team, 2019).

| Leaf litter quantity and quality effects
The consumption by Phylloicus between Cerrado and Amazonia was statistically different (F 1,60 = 4.95; p = .029), as well leaf litter proportions (F 4,60 = 10.85 p < .001), and resource availability (F 1,60 = 8.15; p = .005). We did not record a significant interaction effect between biomes, leaf litter proportions and the resource availability on Phylloicus spp. consumption (Table 1). In Cerrado, F I G U R E 2 Experimental design performed with 80 containers (Cerrado = 40; Amazonia = 40; four replicates per treatment) to assess the effect of litter quality and quantity on the processing by Phylloicus spp. Quality test (a = 100% of Maprounea guianensis; b = 75% of M. guianensis and 25% of Inga laurina; c = 50% of both; d = 25% of M. guianensis and 75% of I. laurina; and e = 100% of I. laurina), in a quality gradient to high (a) to low (e). Quantity test: these five proportions were repeated four for the quantity test, ordered with four leaf disks (Lower quantity) and the second with 12 leaf disks (Higher quantity) in each container higher consumption was recorded in treatment with lower leaf litter quantity (24.0 ± 2.2% standard error) compared with higher quantity (15.0% ±1.7; Figure 3). We recorded higher consumption by Phylloicus sp. larvae in treatments with 100% of M. guianensis (A; 33.0 ± 3.7%).

| D ISCUSS I ON
In general, our results indicated that leaf litter quality were the main factor to explain consumption by Phylloicus larvae, supporting our first hypothesis. In both systems, we recorded higher consumption in treatments with only M. guianensis leaf litter. Leaf litter consumption of M. guianensis in Cerrado (33%), and Amazonia (27%) systems was similar to previous study with this plant species in a Cerrado stream (37%; Rezende et al., 2018). Lignin:N is a worldwide proxy for leaf litter decomposability (Zhang et al., 2019), thus lower Lignin:N and consequently lower hardness explains an accelerated fragmentation by Phylloicus (Biasi et al., 2019). Our results corroborate the knowledge that leaf traits are important for the processing of organic matter Rezende et al., 2014Rezende et al., , 2018.
Therefore, changes in plant species composition of riparian zones and, consequently, leaf litter traits may lead to modifications in litter decay, affecting the functioning of Cerrado and Amazonian streams.
The consumption of I. laurina leaf litter treatment differed between others only in Cerrado, but not in Amazonian system. These differences in consumption of I. laurina were likely due to the greater feeding selectivity of Phylloicus on higher quality litter in Cerrado streams than in Amazonian streams. Recalcitrant compounds (e.g., lignin and cellulose) have a negative influence on colonization of decomposing organisms in Cerrado systems (Casotti, Kiffer, & Moretti, 2014;Rezende et al., 2018). Lower quality litter in Cerrado hinders microbial colonization (Sales et al., 2015), and as consequence affected invertebrate consumption . Furthermore, the low contribution of invertebrate decomposers in tropical streams to leaf litter breakdown is often related to low leaf quality . Added the low palatability of this litter species, the low availability of dissolved nutrients limits the litter decomposition in certain Cerrado streams (Alvim, Medeiros, Rezende, & Gonçalves, 2015;Medeiros, Callisto, Graça, Ferreira, & Gonçalves, 2015). This implies directly in the preference of foliar consumption of Phylloicus spp. by leaf litter with less structural and secondary compounds and higher nutritional content (Moretti et al., 2009;Rezende et al., 2018;Rincón & Martínez, 2006).
On the other hand, our results showed that different mixing proportions between the two leaf litter species had the same consumption response, but unlike when offered separately. These intermediate leaf proportions had the effect of composition of leaf litter mixtures, which may determine decomposition . Different leaf species have different chemical characteristics ; such interspecific differences have substantial effects on leaf breakdown rates (López-Rojo et al., 2018). In a mixture of labile and refractory litter species, shredders focused more on the most labile litter species (Swan & Palmer, 2006) Phylloicus spp. larvae selected softer leaf litter proportions in detriment to less palatable litter for feeding. Although I. laurina has higher nutritional content than M. guianensis, the processing of softer leaf litter appears to be more cost benefit for Phylloicus larvae metabolism (Arias-Real, Menéndez, Abril, Oliva, & Muñoz, 2018), corroborating our second hypothesis. We recorded a higher consumption in treatment with lower quantity compared with higher quantity just in Cerrado system. The higher quantity of resources may cause a smaller stimulus of resources use to satisfy shredders requirements, according to the concept of order of resource selection (Johnson, 1980). Apparently, Phylloicus from Cerrado presents higher stimulus for leaf consumption when there is a smaller litter quantity. Mainly because a vital component to the consumer may be so abundant that the consumer needs only small amounts of it to satisfy their requirement (Johnson, 1980). During periods of less organic matter input in Cerrado streams (Tonin et al., 2017), Phyllocus may have this behavior. The lower consumption of both leaf litter in the Amazonian system may be related to the low quality of them for Phylloicus from Amazonia (Martins, Melo, et al., 2017) highlighting the importance of leaf traits (e.g., C, N and lignin) for the consumption by shredders (Moretti et al., 2009;Rincón & Martínez, 2006)

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study is submitted in the