Abstract
The interpretation of recent numerical results of Shapiro and Teukolsky (1991) concerning the Vlasov equation in general relativity is discussed. The conclusion is reached that, while their computations may well accurately describe the collapse of dust, there are serious grounds to doubt that they are representative of the behaviour of more general solutions of the Vlasov equation. The circumstances under which solutions of the Vlasov-Einstein system may develop singularities are examined and it is shown that there exist initial data not containing trapped surfaces for which this happens. Some comments are made on the significance of these considerations for the cosmic censorship hypothesis.