Around 1980, many perinatal centers began prospective cranial screening of preterm infants using portable ultrasonography at the bedside. This study examined developmental outcome at 1 and 2 years in relation to the presence and severity of neonatal intraventricular hemorrhage (IVH). It varies from earlier reports in the size of the sample, restriction to infants without periventricular leukomalacia (PVL), and an attempt to formulate a predictive model by examining development longitudinally. Parametric and nonparametric analyses demonstrated that IVH related to Bayley mental and motor scores and neurologic ratings at 1 year but not at 2 years. Developmental delay and/or neurologic abnormality were more prevalent in infants with severe IVH but were far from universal. Regression analyses on prediction from neonatal and 1-year performance to 2-year scores revealed significant associations between the 1- and 2-year measures but not the neonatal and outcome measures. A direct insult to the CNS such as IVH thus constitutes only a limited model of risk status.