Correction to ‘Principles of Experimental Design for Ecology and Evolution’

Marshall. D. J. (2024). Principles of experimental design for ecology and evolution. Ecology Letters, 27(4), e14400. https://doi.org/10.1111/ele.14400

In the original publication, under the heading ‘Scales of biology and experiments’, the author made an unjustified assumption that the populations shown in Figure 2 differed solely in pH, the putative causal agent. However, based on the information provided in the original text, the only robust inference was that seeps/upwelling were the likely causal agent. The revised text clarifies this section and presents the argument more precisely.

Revised text:

Scales of biology and experiments

‘As biologists, we recognize that long-term exposures to an environmental stressor might yield evolutionary responses in populations that cannot be predicted from short-term exposures. So, we might use naturally occurring CO₂ seeps or upwelling regions as our driver of low pH; we are using differences among populations to make our inferences. Our factor of interest (lower pH, or more precisely: seep presence/absence) varies at the population level; we therefore need multiple low pH populations and multiple normal pH populations—otherwise, we will be comparing our effect of interest using an inappropriate level of variation (among-individual variation would be used to compare among-population-level differences; Figure 2). Note that we can only infer that pH is the causal factor if no other facet of the populations is different, otherwise, we can infer only that something about the CO₂ seeps/upwelling is likely to drive the differences. The most important point here however is that we need multiple populations of each type, not just one’.

We apologise for this mistake.