Figure 1.

Three possible causal pathways via which behaviour could be associated with telomere length: (a) behaviour causes telomere attrition, (b) telomere attrition causes changes in behaviour and (c) another variable, which we refer to as ‘state’, causes both telomere attrition and changes in behaviour. The two emerging hypotheses for the pattern in the observed data are named according to whether the difference in behaviour is assumed to be causal (causation hypothesis) or secondary (selective adoption hypothesis). (Online version in colour.)