Juvenile survival is one of the least known elements of the life history of many species, in particular snakes. We conducted a mark–recapture study of Crotalus horridus from 1978–2002 in northeastern New York near the northern limits of the species' range. We marked 588 neonates and estimated annual age-, sex-, and morph-specific recapture and survival rates using the Cormack-Jolly-Seber (CJS) model. Wild-caught neonates (field-born, n  =  407) and neonates produced by captive-held gravid females (lab-born, n  =  181) allowed comparison of the birthplace, or lab treatment effect, in estimated survival. Recapture rates declined from about 10–20% over time while increasing from young to older age classes. Estimated survival rates (S ± 1 SE) in the first year were significantly higher among field-born (black morph: S  =  0.773 ± 0.203; yellow morph: S  =  0.531 ± 0.104) than among lab-born snakes (black morph: S  =  0.411 ± 0.131; yellow morph: S  =  0.301 ± 0.081). Lower birth weights combined with a lack of field exposure until release apparently contributed to the lower survival rate of lab-born snakes. Subsequent survival estimates for 2–4-yr-old snakes were S  =  0.845 ± 0.084 for the black morph and S  =  0.999 (SE not available) for the yellow morph, and for ≥5-yr-old snakes S  =  0.958 ± 0.039 (black morph) and S  =  0.822 ± 0.034 (yellow morph). The most parsimonious model overall contained an independent time trend for survival of each age, morph, and lab-treatment group. For snakes of the first two age groups (ages 1 yr and 2–4 yr), survival tended to decline over the years for both morphs, while for adult snakes (5 yr and older), survival was constant or even slightly increased. Our data on survival and recapture are among the first rigorous estimates of these parameters in a rattlesnake and among the few yet available for any viperid snake. These data are useful for analyses of the life-history strategy, population dynamics, and conservation of this long-lived snake.