Relationships between traits and fitness are defined within a context, but the context varies continuously
Selection coefficients and gradients measure the slope of the relationship between the realized values of a trait and the relative fitness (that is, individual fitness divided by average population fitness) of each individual within a population. But these trait-fitness relationships, or phenotypic selective pressures, make sense only in a given environmental context.
However, environmental contexts vary endlessly. The environment that affects a trait-fitness relationship is complex and acts at different levels. First, at the within-individual environment level (the phenotypic environment), values of a trait would have different fitness outputs depending on the values of other characters within the integrated phenotypes. These different fitness outputs may occur even when the abiotic and biotic environment is equal. This is due to the existence of indirect and correlated selection, as shown above.
Second, at the population environment level, a single trait may be related to different fitness outputs where trait values can be positive, neutral or negative depending on population parameters such as population density (see for example Yang et al 2018).
Third, at the community level, due to mutualistic and antagonistic interactions between species, the expected effects of traits on community composition loop back to modify the relationships between traits and fitness.
Last, at the ecosystem level, the abiotic and biotic environments continuously vary, spatially and temporally. These variations are particularly important, regarding the potential unpredictable shifts driven by Global Change.
Given all the direct and indirect sources of selection previously described, and because of the complexity laid out here of the selective pressures over the integrated phenotype and the varying environment, it is reasonable to argue that any trait would be related to fitness in at least one of the potential environments. That is, traits would be related to fitness depending on environmental variation. For example, a plant population would undergo strong selection on root traits during a drought year, and the same plant population would undergo strong selection on flower traits during a year with severe limitation of pollinators. Nonetheless, if we study the same population during a year with optimum precipitation levels and plenty of pollinators, we cannot argue that root traits or flower traits in this species are non-functional traits.
If in any one of all possible environments a potential link exists between the distribution of values of a trait and fitness, then all traits are potentially linked to fitness, and thus are functional by definition. Evolution happens through ecology, and the object of study in ecology -diversity at different levels- is the consequence of evolution. We cannot rend one from the other. Rather, ecology and evolutionary biology should be integrated in the study of causes and consequences of trait diversity, a possible start point is to acknowledge that all traits are functional.
Acknowledgement s Thanks to R. Lande, M. Kulbaba, C.M. Herrera and L. Sampedro for comments on the manuscript.
Data accessibility statement. NA
Competing Interests. Since this paper explains why the use of the term “functional trait” is redundant, and even misleading,we can guess that scientists working in the area of “functional traits” would disagree with the publication of the arguments here explained.
Author contributions. NA
Funding information. Mar Sobral was benficiary of a grant from Xunta de Galicia I2C (Spain) when performing this work.