Figure 1.

Simulations of probability of population persistence for 10 000 replicates as a function of behavioural responses (Neutral, random behavioural responses; GoodChoice, matching habitat choice; BadChoice, habitat mismatching choice; ReprSkip, reproductive skipping; LearnExpl, learning through exploration; LearnBreed, learning from breeding experience) for different initial population sizes according to different life histories (fast and slow). Simulations have been run with the same deterministic growth rate (λ) of 1.05 and moderate behavioural responses, under the five different scenarios: phenotype–environmental matching (scenario 1) and phenotype–environmental mismatch causing moderate increases of adult mortality (scenario 2.1), extremely high adult mortality (scenario 2.2), moderate increases of juvenile mortality (scenario 3.1) and extremely high juvenile mortality (scenario 3.2). Simulations with strong behavioural responses are shown in the electronic supplementary material, figure S2. The fast strategy is characterized by early onset of first reproduction (1 year old), high annual fecundity (q = 8) and low adult survival ($p_{1,s_{\mathrm{b}}}=0.4$), while the slow strategy exhibits delayed onset of reproduction (3 years old), low fecundity (q = 8) and delayed onset of first reproduction but high adult survival ($p_{1,s_{\mathrm{b}}}=0.85$ ). Note that in scenario 1, the two habitats are the same, and therefore, all behavioural responses except reproductive skip are equivalent to the neutral behaviour.