Test–retest reliabilities and practice effects of a broad range of neuropsychological measures were examined in 384 normal or neurologically stable adults. Median test–retest interval was 11 months (range 3–16 months). The reliability estimates for most of the measures are reasonably good, ranging from .70 to low .90s. An exception is the relatively poor reliabilities of most memory measures. For all test measures, the value on initial testing is a strong determinant of the value on the second examination. Practice effects are seen on most measures. The magnitude of the practice effects, however, varies as a function of type of measure, test–retest interval, age, and overall competency level of the participant. This study provides several types of retest information that may be useful for future research and clinical work: comparative reliabilities of the various measures, estimate of error variability associated with each administration, standard deviation of the change, and comparative magnitude of practice effects on various tests. (JINS, 1999, 5, 346–356.)