Analysis of the pulsar P-P distribution

A new technique for the systematic evaluation of models of pulsar period evolution on the basis of a complete observational sample is outlined and applied to the existing incomplete sample. Possible selection effects are discussed. It is concluded that the simple law for the rate of change of period P, dP/dt proportional to Psup(2-n) is incompatible with the assumption of stationarity and pulsar 'death' at large periods if n > 2. Models with n < 2, or with n approximately > 2.5 and torque decay on a time-scale of 1 Myr are consistent with the data. Another possibility is that the beaming fraction decreases by a factor approximately 5 as a pulsar slows down. A new procedure for deriving a rigorous lower limit to the creation rate of pulsars within the sample is presented, and it is shown that most pulsars appear to be born with large values of dP/dt. (author)