This thesis focuses on different aspects of radio pulsars: on the one hand it describes research on the pulsar emission mechanism (nulling, drifting subpulses, single-pulse polarisation and their mutual interaction). On the other it gives an account of investigations into the evolution of the pulsar population as a whole (birthrates, velocities, magnetic field evolution) and possible ways to probe this population (simulations of upcoming pulsar surveys). In Chapters 1–3 we investigate observationally how subpulse drift behaves after the ‘nulls’ (stops of all emission) that some pulsars show. We find that the drifting-nulling interaction tells much about the true motion of the sparks; the spread in the inferred velocities of the sparks is much larger than previously thought possible. In Chapters 4–6 we study the interaction of drifting subpulses, refraction and polarisation. We show how refraction changes subpulse patterns (Chapter 4); we find that the large polarisation-angle changes observed within single subpulses can be explained in terms of orthogonal polarisation modes (Chapter 5); and we demonstrate that, in contrast to the handedness, the amount of circular polarisation in pulse profiles is independent of frequency (Chapter 6). In Chapter 7 we focus on pulsar evolution, and show that magnetic field decay is not needed to explain the observed relative lack of long-period pulsars; furthermore, we find that the scarcity of high-field radio pulsars does not have to mean that high-field pulsars (‘magnetars’) do not emit in radio: even if many are born, their evolution is so fast that only few, if any, are visible at a given time. In Chapter 8 we extend this population synthesis to LOFAR. We simulate and compare strategies for finding new pulsars and find that a 60-day all-sky survey can find 1500 new radio pulsars, doubling the currently known population.