Many writers have attempted to develop criteria to demarcate between competent science and pseudo?science. Such attempts can be aimed at sizeable, organized endeavours, such as mesmerism and astrology, or at the level of individual practice. The latter is seen by some, such as Lugg, as more likely to be feasible and useful. This paper argues that parapsychology, due to its complexity and diversity, illustrates some of the problems of attempting to develop demarcation criteria for extensive endeavours. It is also suggested (...) that parapsychology may offer a productive ground for testing whether demarcation criteria can be successfully applied to practices in respect of predicting which will succeed and which will not. The conclusion is that, for demarcation efforts to be useful, they should pass some of their own criteria, such as falsifiability. (shrink)