Why body size matters: how larger fish ontogeny shapes ecological network topology

Abstract

Ontogenetic development can strongly shape species interactions. Yet, rarely is stagestructure considered when analyzing species interaction networks, particularly networks that can account for more than feeding relationships. Here, we assess 1) if body size or trophic level regulate the importance of species' ontogeny on their interactions and 2) how including relevant stage-structure affects the topology of species interaction networks. We use a count-based inferential method to create networks from adult and juvenile fish count data and test stage-structure importance by comparing a model that includes stage-structure for all species against models that include stage-structure only for larger fishes and only for piscivorous fishes during network construction. While the inferential method we use cannot differentiate between different types of interactions, it can account for different interaction types within a network as a pairwise interaction is inferred when one species influences the abundance of another. Next, we use graphlet-based techniques to test if including stage-structure alters overall network topology and a linear model to measure if adult-juvenile size differences drive interaction differences at a species-level. We find that the model that includes stagestructure only for larger fishes outperforms other stage-structured models including the model with only piscivore stage-structure, and that larger differences in body size among juveniles and adults lead to greater interaction dissimilarities. Moreover, we find topological differences between inferred networks that only include adults and those that account for the stage-structure of larger species. Overall, our study demonstrates how stage-structured topological changes can be measured using inferred interaction networks and illustrates how larger species' juveniles fundamentally shape the structure of stream fish communities.

How network size strongly determines trophic specialisation: A technical comment on Luna et al. (2022)

“…One major limitation from Luna et al (2022), however, is their use of open networks without appropriate controls for non‐systematic sampling and differences in network construction. Without these controls, networks likely contain structural differences due to, for example, differences in the amount of sampling time (CaraDonna et al, 2021), sampled area (Galiana et al, 2018) or from differences in node resolutions (Bodner et al, 2022; Hemprich‐Bennett et al, 2021). While these differences can prevent commensurability and therefore should be appropriately identified and controlled (Jordano, 2016), details about open networks are often unavailable, forcing researchers to rely on other approaches to account for these structural differences.…”

No strong evidence that modularity, specialization or nestedness are linked to seasonal climatic variability in bipartite networks

“…While some traits may be generalizable across organisms like body size, finding other common traits that govern different organisms, for example, both insect and hummingbird interactions within plant–pollinator networks, may not be easy. Moreover, the life stage resolution of nodes can contribute to significant differences in network structure (Bodner et al, 2022; Clegg et al, 2018). For example, separate nodes are likely required to represent different life stages of species for those in which life stages act and behave entirely differently from one another, for example, tadpole and frog, resulting in even greater difficulties when comparing with networks without such variations between ontogentic stages.…”

Shortcomings of reusing species interaction networks created by different sets of researchers

“…Hemprich-Bennett and colleagues (2021) [39], Bodner and colleagues (2022) [40] https://doi.org/10.1371/journal.pbio.3002068.t001…”