OBJECTIVE: Our goal was to describe the variability of sleep duration (time in bed per 24 hours) in healthy children from 1 to 10 years of age in comparison with growth measures. METHODS: A total of 305 children were followed with structured sleep-related interviews and measurements of height and weight 12, 18, and 24 months after birth and then at annual intervals until 10 years of age. SD scores were calculated, and smooth curves were fitted by smoothing splines through the SD scores. The long-term variability channel within children (units SD score) was defined as the difference between the maximum and the minimum of the smooth curves and the short-term variability channel (units SD score) as the difference of the largest and the smallest deviations of the original SD scores from the smooth curve. RESULTS: Sleep duration remained within a long-term variability channel <0.5 SD score in 21% of the children (34% for height, 21% for weight). Nearly every second child (46%) stayed within a long-term variability channel <1.0 SD score (76% for height, 64% for weight). Sleep duration of approximately 90% of all children ran within a long-term variability channel of <2.0 SD score (corresponding, eg, to the range between the 2nd and the 50th percentile). No single child's sleep duration remained within a short-term variability channel <0.5 SD score, indicating fluctuations from year to year (60% for height, 53% for weight). An association between aspects of sleep duration and somatic growth was not observed at any age. CONCLUSIONS: Sleep duration during early and middle childhood shows large variability among children, as well as trait-like long-term stability and state-like yearly fluctuations within children. An individual approach to the child's sleep behavior is needed; expectations in terms of appropriate sleep duration of the child should be adjusted to the individual sleep need.