Environmental Drivers and Network Structure of Hylid Anurans in Floating Meadows From Amazonian Oxbow Lakes

23 Despite the increasing amount of knowledge available regarding the ecological interactions between species, the 24 dynamics of anurans in aquatic environments are little explored and understood. In this way, our work aims to 25 assess which factors influence the composition and the ecological interactions of hylid anurans in oxbow lakes 26 in the middle Purus River, Amazonas. We sampled three lakes with high, medium and low levels of 27 connectivity twice, once during the flood and then in drought hydrological regimes. Variations in the hylid 28 anuran assemblages and ecological interactions were tested as function of environmental niche, food resources, 29 level of connectivity and hydrological regime. The availability of environmental resources and the availability 30


43
In aquatic environments, changes in community structure are attributed to the sum of multiple processes,

76
The use of tools from the Theory of Complex Networks has been useful to understand new aspects of 77 the ecology and evolution of species assemblages. This is because these tools allow us to study how the species still scarce. Indeed, we still lack theory and natural history to understand how the anuran assemblages interact 86 with macrophytes assemblages in floating meadows (i.e., as an environmental resource for anurans) and their 87 associated macroinvertebrates (i.e., as a part of the diet for anurans). Therefore, investigating how interactions 88 are affected by changes in river-lake connectivity and hydrological regimes will increase of understanding of 89 how freshwater communities are assembled and maintained.

4
The aim of this study was to examine how the communities of hylid anurans can vary with macrophyte 91 assemblages in floating meadows and their associated macroinvertebrates, and how these interactions are 92 affected by changes in river-lake connectivity and hydrological cycle. Specifically, we tested the following 93 hypotheses: (i) the environmental niche and available food resources are the primary predictors explaining the 94 variation in composition of hylid anurans, although it should be dependent on both river-lake connectivity and  cycle and lake-river connectivity, are important components to structure anuran communities, (ii) we expect that 103 interactions between hylid anurans, macrophytes and macroinvertebrates vary in function of river-lake 104 connectivity and hydrological regime. To this end, we identified the macrophyte species where each specimen 105 was registered and analyzed its stomach content. According to the optimal foraging theory (OFT), when food is 106 scarce, individuals expand their diet width to include less-favored items, which may expend their trophic niches 107 and therefore affect their ecological networks (Emlen 1966;Robinson and Wilson, 1998). Thus, we predict that 108 the oxbow lake more isolated and with low connectivity level has a high number of interactions (i.e., high 109 connectance), in response to the lower supply of food, resulting in non-restrictive diets. In contrast, because 110 more connected lakes have higher migration taxes (e.g., macrophyte species; Mormul et al. 2013), they can 111 result in a great prey availability, leading to a maximum of prey selectivity and more specialized networks. We 112 expect also that interactions between anurans and invertebrates would display modularity as it has been shown 113 that such networks are not nested, but they show low levels of specialization and modularity and high    (Table 1). Thus, our observations in the field allowed us to infer that the connectivity 154 profile (variation of terrain elevation) is directly related to the degree of connectivity. For example, although the 155 high connectivity lake has longer connectivity and depth of connectivity similar to that of medium connectivity, 156 the lower profile of the terrain along connectivity allows it to remain connected with the river during all 157 hydrological regimes.

193
The stomachs of the collected specimens were removed, the contents were scattered in a Petri dish, counted and

268
During the flood, the more abundant macrophyte species were P. stratiotes in lakes of high and medium 269 connectivity (42% and 35%, respectively) and E. crassipes (34%) in low connectivity. During the hydrological 270 regime of drought, the more abundant species were P. stratiotes (25%) in lakes of high connectivity, E.

11
We sampled 4041 individuals of macroinvertebrates. Individuals not identified or in the larval stage

273
were not used in the statistical analyzes. The more abundant invertebrate taxons were Hemiptera (18%),

308
Network metrics according to regime, connectivity and by the interaction type are shown in Table 2.

309
Connectance showed an increasing trend from high to low connectivity lakes in both interaction types. The

337
Our results support our first hypothesis, by showing that the availability of environmental resources (i.e., species 338 of macrophytes) and the availability of food resources (i.e., macroinvertebrates) are the best factors that explain 339 the distribution of hylid anurans. In addition, a high percentage of explanation indicates that the variation in 340 abundance of hylid anurans is highly dependent on the variation in the macrophytes and macroinvertebrates 341 composition between the hydrological regimes. Moreover, our results support the second hypothesis as we 342 found that the connectance, modularity and specialization of anuran, macrophytes and macroinvertebrates 343 networks varies depending on the connectivity and the hydrological regime.

344
We found that the abundance of the hylid species is influenced by all the factors that we measured (i.e., 345 availability of environment, connectivity, food resource and hydrological regime). However, the availability of 346 the environmental and food resources was the best factor explaining the variation in the abundance of anuran

400
The primary prey of anurans is insects, as well as other arthropods, such as spiders and mites (Simon

433
Conflicts of interest/Competing interests -Authors declare that they have no conflict of interest.

435
Availability of data and materials -All data produced from this study are provided in this manuscript.