Apparent competition leaves no detectable imprint on patterns of community composition: observations from a natural experiment

Abstract

Indirect interactions mediated by natural enemies shared among herbivorous insects have recently attracted much interest. While many studies have predicted a high potential for apparent competition, only a few have rigorously tested predictions derived from the food web structure in terms of realised population and community dynamics. In this study, a quantified food web was used to identify pairs of herbivore species potentially tied by strong parasitoid‐mediated interactions. The host populations and their parasitism rates over two consecutive years were then followed, during which time the abundance of one dominant host crashed. Following this natural experiment, imprints of asymmetrical, long‐term apparent competition among hosts were sought in population growth rates. The population growth revealed no signs of apparent competition as mediated by parasitoids: the abundance and parasitism rates of less abundant herbivores were uncorrelated with the relative growth rate of four dominant herbivore species with which they shared a major part of their parasitoids. Likewise, the population crash of the dominant herbivore, Cynips longiventris, did not detectably affect the abundances or parasitism rates of other herbivores with which it shared a major fraction of its parasitoids. Overall, this case study fails to add evidence to previous suggestions that apparent competition may constitute a major force in structuring natural communities. However, it suggests that dynamics should not be inferred from static food web structure alone, but rather that predictions based on structure should be verified by empirical observations of realised population dynamics.

Related herbivore species show similar temporal dynamics

“…Despite an overall imprint of phylogenetic relatedness, species with more similar parasitoid communities did not exhibit any more similar (or dissimilar) temporal dynamics than species attacked by distinct parasitoid assemblages. This is consistent with earlier work at the Finnish study site, where a previous multi-year experiment (Tack et al, 2011), as well as a natural experiment (Kaartinen & Roslin, 2013), failed to reveal any detectable impact of increased densities of herbivore species i in year t − 1 on the abundance of herbivore species j (sharing parasitoids with herbivore species i) in year t. This contrasts with empirical field studies showing apparent competition in plant-feeding insects (Blitzer & Welter, 2011;Frost et al, 2016;Morris, Lewis, & Godfray, 2004). Indeed, while parasitism usually causes high mortality in plant feeding insects (Hawkins, Cornell, & Hochberg, 1997), studies have generally failed to find an impact of parasitoids on the pattern of herbivore cycling and outbreak dynamics (Hagen, Jepsen, Schott, & Ims, 2010;Schott, Hagen, Ims, & Yoccoz, 2010).…”

Species–area relationships across four trophic levels – decreasing island size truncates food chains

“…In addition to effects on emergent properties such as food chain length, island size may also affect the interaction structure among species. While spatially explicit data on who feeds on whom within habitat islands may be derived for specific systems or sites (Kaartinen and Roslin 2011, 2012, 2013), such data are practically impossible to obtain for the current system. Thus, our inferences regarding food chain length were built on patterns of co‐occurrence, not on trophic interactions resolved at the level of species pairs (see above under Plant and insect data).…”

How to make more out of community data? A conceptual framework and its implementation as models and software

“…Linking species co-occurrences to traits, phylogenies, and their responses to environmental variation would provide new tools for addressing several intriguing questions, e.g. whether species with few strong interactions are more vulnerable to disturbances than species with many weak interactions (Fortuna & Bascompte 2006;Aizen et al 2012;Abrego et al 2017b), or whether the extent to which prey species within communities are coupled by predators translate into correlated abundances in space or time (Morris et al 2004;Tack et al 2011;Kaartinen & Roslin 2013). We also note that the estimated association networks are conditional on species occurrence: even if two species show a strong positive co-occurrence, the association is not realised in areas where neither of the species occurs.…”