Nine neonatal Long-Evans rats had continuous (24 h/day) polysomnography for 2 weeks, from age 14 days through age 27 days. A new finding was that six more or less independent measures of REM sleep occurrence decreased in parallel from age 14 days to age 27 days. The measures included parallel decreases of four measures of 24-h REM duration (tonic REM sleep, phasic REM sleep, mean REM period duration, and number of REM periods) along with parallel increases of two measures of REM delay (REM latency and percent of nonsleep onset REM periods). A parsimonious interpretation of the correlated changes is that a common developmental REM sleep inhibitory process accounts for the six parallel changes over time. This hypothesis can be tested empirically by studying inhibitory processes that operate on the pedunculopontine tegmental/latero-dorsal tegmental nuclei, the generators of REM sleep. The study also noted that compared with (same species) normal adults, endogenous depressives had the same distinctive REM sleep characteristics as neonatal rats. The similarity suggests that an underdeveloped, relatively weak REM sleep inhibitory process may account for the REM sleep peculiarities of endogenous depression. This hypothesis can be tested in adult rats made "depressed" by neonatal treatment with antidepressant drugs. Thus, the ontogeny of REM sleep suggests a developmental process that may be altered in humans predisposed to endogenous depression, and may account for the (life-long) REM sleep abnormalities of the disorder.