The objective of this study was to establish longitudinal normative limits for home memory monitors during early infancy. Eighty-eight healthy infants were monitored overnight at 0.25-19 wk of age using the Healthdyne Smart Monitor. Apnea settings were 14 s for recording and 40 s for alarm; the bradycardia setting was 50 beats/min (5-s delay) for both recording and alarm. Arterial oxygen saturation (SaO2) was documented whenever an event was recorded. The monitor was used 77% of all possible days; median daily use was 8.0 h. Eighty-three percent of all monitor alarms were caused by loose leads, the other 17% by false apnea or false bradycardia. Of all recorded events, 68.9% were caused by false apnea or false bradycardia; the other 31.1% were central apneas that reached the recording threshold of 14 s. The longest apnea was 36 s (wk 1); the 95th percentile for longest apnea was 19.9 s in wk 1 and 18.0 s in wk 17-19 (p < 0.001). Periodic low SaO2 values occurred with periodic breathing; the lowest value was 72%. The 5th percentiles for lowest SaO2 were 82 and 86% in wk 1 and 13-19, respectively (p < 0.001), but the minimum value observed in any week was never > 81%. The median duration of SaO2 < 90% was only 5 s but the range was wide (1-183 s), and 39/527 episodes (7.4%) were > 10 s. In summary, these longitudinal data provide the first available normal limits for cardiorespiratory pattern and SaO2 during documented home monitoring in early infancy. Utilization of these normative data will improve the diagnostic validity and clinical usefulness of event recordings.