The paper discusses the regularity account of causation but finds it insufficient as a complete account of our notion of causality. The attractiveness of the regularity account is its attempt to understand causation in terms of empirically accessible features of the world. However, this account does not match our intuition that singular causality is prior in normal epistemic situations and that there is more to causation than mere succession. Apart from succession and regularity, the concept of causality also contains a modal feature which allows us to engage in counterfactual discourses about singular causal events and to claim that a particular cause is both sufficient and necessary for its effect in the circumstances. However, we may directly observe singular causes, but the modal element is not something we can possibly observe. Rather, this element is something we add to our perception of succession. Thus, the paper suggests that the modal feature of causality is a mental construction which was originally formed by our knowledge of certain structural features of similar events in other situations. It stems not from what we actually observe but from what we have observed or may observe under different but relevant circumstances. So the concept of causation has partly an empirical content and partly a constructed one.