I disagree with your advice to stop searching for universal principles in ecology (Nature 507, 139–140; 2014). It is fortunate that Darwin held no such view. And in this era of big data, big theories are still needed — even in ecology.

The “caveats and exceptions” in existing theory, which you imply scuttle efforts to build better theory, are the very foundation for improvement. This is how science progresses; ecology is no exception.

You recommend that “ecologists should embrace the non-predictive side of their science”, but prediction is part of what distinguishes science from other worthwhile endeavours; it enables us to strengthen and refine our understanding.

Using the wolves in Yellowstone National Park as an example to support your view is ironic. Top-down trophic control in multispecies populations is not a theory: it is an oversimplified conceptual model — a good story that is known to apply sometimes. It remains to be seen whether a unified theory of trophic, competitive and facilitative interactions among species can be achieved, but ecologists should not be advised to give up now.