Over one million individual racing records were used to describe racing time as a measure of racing performance in American Quarter Horses. Contemporary groups were defined, adjustments for sex, age and handicap weights were calculated and estimates of heritability and repeatability were computed. Five racing distances (201, 320, 366, 401 and 796 m) were represented and racing times were linearly related to distance. Frequency distributions of racing time were skewed and more peaked than a normal distribution. Repeated records were an important source of information for increasing the accuracy of a genetic evaluation. Selection, particularly selection involving horses that are to remain in the racing circuit, was evident in Quarter Horse racing data. Mean racing time in geldings improved with age, relative to stallions and mares, possibly because the faster stallions and mares were used for breeding stock while the faster geldings continued to run. Stallions were faster than mares at all ages but geldings ranked differently at various ages. Racing performance in Quarter Horses improved until horses were four or five years of age and declined after age six. Handicap weight had only a small effect on racing performance and was not considered a major source of environmental influence. Individual races within tracks, years and days controlled a major portion of the variation in racing time (over 60%) and were defined as contemporary groups within which genetic differences among horses were evaluated. Heritability and repeatability estimates for racing performance were data dependent. Higher estimates were obstained from two-year-old horses than from records of older horses. Lower estimates were obtained when a single track was used as the data source than when all records were used. Heritability estimates were extremely variable and ranged from -.29 to.38 while repeatability estimates were more consistent (.20 to.51). Overall weighted means for estimates of heritability and repeatability based on an analysis of variance were.22 and.32.