Our main study species expresses polymorphic colouration between and within populations despite strong frequency-dependent selection on warning signals (see Lindström et al 2001).

We have discovered that this colour polymorphism is an outcome of multiple, sometimes conflicting, selection pressures. 

For example, our experiments show that, in Scandinavia, yellow morphs are better protected against avian predators but white morphs have better mating success (Nokelainen et al. 2012). Attack risk can also vary between morphs depending on the predator community composition: yellow males are better protected when Paridae (e.g. great tits) are abundant, whereas white males have the advantage when Prunellidae (dunnocks) are abundant (Nokelainen et al. 2014). 

Furthermore, there are trade-offs between warning signal efficacy and thermoregulation (Hegna et al. 2013). Not to mention genetic correlations between colouration and immunity mechanisms (Nokelainen et al. 2013).

Taken together, our results are indicative of a more complex dynamic between predators and aposematic prey than suggested by classic views of positive frequency-dependent selection favouring local morphs.