The inter- and intra-observer agreement for the assessment of several aspects of the quality of spontaneous movements in newborn infants was analysed by 4 observers. The agreement rates were corrected for chance by the use of kappa statistics, which revealed kappa values ranging from 0.36 to 0.84. The degree of agreement on various items differed. A moderate-to-good agreement was found for the items "onset of movement," "speed," "speed of arms compared to legs," "amplitude of arms compared to legs," "fluency," "variability in arm patterns" and "global judgement of the movement" (kappa values ranging from 0.51 to 0.84). Less agreement occurred in the items "amplitude," "variability in movement pattern" and "variability in leg patterns" (kappa values ranging from 0.36 to 0.40). Even though it is believed that the qualitative assessment of spontaneous motor behaviour is very subjective, our results suggest that several qualitative aspects of infant motor behaviour can be reliably judged. We concluded that the global judgement of the quality of general movements in full-term infants is a reproducible, and therefore reliable, clinical tool for evaluating the function of the central nervous system in the neonatal period.