The movement and posture abnormalities that are characteristic of cerebral palsy are associated with underlying abnormalities of muscle tone, including dystonia, spasticity, and rigidity. The subtypes of cerebral palsy are defined according to the predominant muscle tone abnormality, its distribution and severity. There is widespread agreement that CP subtypes should be divided into the following groups. Spastic subtypes (hemiplegia: unilateral asymmetric spasticity; diplegia: bilateral symmetric spasticity, lower limbs more affected than upper limbs; quadriplegia: bilateral symmetric spasticity, upper and lower limbs affected equally), dystonic or athetotic type; and other miscellaneous types: hypotonic, mixed types, etc. These descriptive definitions are not useful when forming homogeneous cohorts of CP subtypes in rehabilitation research studies. Even when a preceding training session establishes precise criteria for each subtype based on these definitions and uses a "four limb" approach (defining the dominant muscle tone abnormality and its severity in each limb) the interobserver agreement in the classification of CP subtypes remains poor. This may be because there are no widely accepted definitions for the most common muscle tone abnormalities in cerebral palsy (spasticity, rigidity, dystonia) and examination methods vary widely. Recently, an interdisciplinary group of clinicians published a consensus document in which they set forth clear definitions for these disorders of muscle tone and recommendations on how to examine for them. Since the CP subtypes are defined according to the dominant muscle tone abnormality, we hypothesize that it may be possible to achieve acceptable interobserver agreement in the classification of the CP subtypes if these definitions and standard examination techniques are incorporated into the preceding training session with the other accepted elements. To test this hypothesis we will train clinicians from three different areas of expertise: child neurology, physiatry, physical therapy to classify subtypes of CP using accepted methods (establishing precise criteria for the subtypes and using a four limb approach) and will also train them to identify muscle tone abnormalities using the definitions and standard examination techniques in the consensus document. Following this training, we will ask the same clinicians to classify subtypes of cerebral palsy in a group of CP children. We will then measure interobserver agreement for the classification of CP using the kappa statistic. Achieving acceptable interobserver agreement in the classification of CP subtypes may facilitate the formation of homogeneous cohorts in research studies.