Wild horses (Equus caballus) at Pryor Mountain were studied by direct observation from 1993 through 2007. All horses present were individually identifiable on the basis of coat coloration, head and leg markings, gender, and band associations. Of the 609 horses either present prior to foaling in 1993 or born since, ages were precisely known for 491 (observed as a foal). Ages for 52 horses were estimated through tooth eruption and wear patterns, and for the remaining 66 horses through body size, morphology, and anecdotal evidence concerning when they were present on the range. At varying intensities, never less than 30 days per year, all horses were inventoried and their band associations noted. Foals were paired with dams based on observations of attachment during the early days and weeks of life. Year of death was determined by identification of the carcass where possible. In the absence of finding a carcass, an animal that was not observed for 2 years was considered to have died in the year that it went missing. Animals that were removed from the herd and mares that were part of a contraception study were excluded from calculations of survival and foaling rates, respectively, as appropriate. The average prefoaling population over the 15 years of the study was 148.8 animals (range = 120-187), and the annual foal crop averaged 32.1 (range = 23-40). Large removals (19-60 animals) in four years helped maintain the herd at this level; apparent growth rate (calculated as though removals had not occurred) was 9.6 percent annually (? = 1.096, range = 0.977-1.220). This annual growth rate is relatively low compared to that for many western horse herds, at least in part because of a decline in foal survival. Sex ratio of the foal crop varied widely among years, but pooled across years did not differ from 50:50. Sex ratio in the herd changed mostly as a result of removals. The average age of both males and females in the herd increased during the course of the study. Annual survival of males did not differ from that of females, nor did gender affect annual survival of foals. Pooled across years, ages, and sexes, the annual survival rate was 0.899. Annual foal survival rate was 0.697 and declined through time, with a tendency toward recovery in 2005-2007. Foal survival was higher in larger bands, but did not differ between foals born to primiparous and multiparous mares. A few 2-year-old mares produced foals; foaling rate (excluding contracepted mares and foals they produced) increased through age 10, remained high through age 15, and declined thereafter. Overall foaling rate for mares =3 years of age was 0.576 foals per mare, with no apparent trend during the period of our study. Foaling rate in years following gathers was somewhat lower than in other years. There was a positive relation between foaling rate and band size. Primiparous mares were somewhat less likely to foal in the following year than were multiparous mares. Most stallions that acquired a harem did so at age 5 or 6, and the average age of harem stallions increased during our study. Most harems had 1-3 mares =2 years of age, but harem size varied with age of the stallion, increasing through about age 11 and declining thereafter. About 6 percent of bands had a satellite stallion (=5 years of age), but the mean number of mares did not differ between single- and multistallion bands. Most stallions left their natal band at age 2 or 3, but 17 percent remained with their natal band until age 4 or 5. Foal survival rate was positively related to precipitation, suggesting a possible link to forage production and availability mediated through mare fitness. There also was evidence for density-dependent population regulation, as both population growth rate and survival rate were negatively correlated with population size from the previous year. These and other factors were not sufficient to stabilize the population during our period of study, however, as evidenced by the necess