Parents may respond in various ways to the birth of a seriously affected child. They may, for example, decide not to have any more children or to have one more child and then stop. These various responses are called "stoppage rules" in the genetic literature. Where stoppage rules are operative, the order in which affected and nonaffected children are born is disturbed in definite ways. The present paper shows that stoppage rules are at work in a recently reported data set consisting of 46 multiplex families of childhood autism and, as a consequence, that the segregation ratio was underestimated in the original report. The implications of these results for genetic studies of autism are then discussed.