Cerebral palsy (CP) is a term of convenience applied to a group of motor disorders of central origin defined by clinical description. It is not a diagnosis in that its application infers nothing about pathology, aetiology, or prognosis. It is an umbrella term covering a wide range of cerebral disorders which result in childhood motor impairment. The precise inclusion criteria vary with the objectives for using the term. For meaningful comparison of rates of CP, as performed by and between CP registers, it is important that the rates should be generated using the same criteria. As generally understood there must be motor impairment, and this impairment must stem from a malfunction of the brain (rather than spinal cord or muscles). Furthermore, the brain malfunction must be non-progressive and it must be manifest early in life. For the purposes of comparisons of rates across time even when the condition meets all the above criteria, it must not historically have been excluded from the category of CP. This paper addresses the problem of standardizing the inclusion criteria for selecting people included on CP registers with particular reference to this last criterion.