In a seminal contribution published in 1972 (Nature, 238; 413-414), Sir Robert May showed that from a mathematical point of view the more complex an ecological community is (in terms of the number of species and interactions in the system), the less stable it is. However, complex ecological communities are observed in nature, and so the issue on how species in large complex ecological communities may coexist is still a relevant and open debate in ecology.
In recent years several searched for new principles allowing ecosystems to persist despite their complexity, but a general consensus on this topic has not yet been achieved.
Last summer, an intriguing work published in Science (A. Mougi, M. Kondoh, Science 337, 349) claimed that specific mixture of antagonistic (predator-prey) and mutualistic interactions (beneficial for the interacting individuals) between species is likely to contribute to stabilize ecological communities. Furthermore, they also found that in this type of hybrid community “…increasing complexity leads to increased stability”. As mixing of interaction type is the norm rather than the exception in ecological communities, these conclusions might have led to a final word in the “complexity-stability paradox”.
In our work “Disentangling the effect of hybrid interactions and of the constant effort hypothesis on ecological community stability”, published Early View in Oikos, we show that this is not the case. Indeed, we proved that mixing of mutualistic and predator-prey interaction types does not stabilize the community dynamics and we demonstrate that a positive correlation between complexity and stability is observed only if species interact so that generalist species (the ones with several “partners”) interact very weakly (in terms of intensity) with respect to specialist species (which have only few partners). We also show that the main findings presented in Mougi and Kondoh work arise as an artifact of the peculiar rescaling of the interaction strengths they imposed. Indeed, using their methodology, the very same effect of ecosystem stabilization may be obtained for generic random ecological networks.
In conclusion, the mismatch between theoretical results and empirical evidences on the stability of ecological community is still there also for communities with a mixing of interaction types, and the “complexity-stability” paradox is still alive. Our work suggest that complexity and stability may be reconciled if a particular scaling of the interactions strength with the species degrees (number of resources) exists, but further studies and experimental evidences are still needed to better understand the role of interaction strengths in real ecological communities.
Samir Suweis, Jacopo Grilli, Amos Maritan