The time devoted to eating, vigilance (or scanning), walking and aggression was studied in one-minute focal animal samples as a function of sex, age, and motherhood for several African animals including African elephant, Burchell's zebra, Cape buffalo, Defassa waterbuck, impala, and Uganda kob. Between 24 and 45 percent of the variation in vigilance behavior was accounted for by location (center or edge of herd), sex, age and whether a female had young. Time of day entered the linear multiple regression models for two species, and surprisingly group size did not enter any of the linear multiple regression models in 1993. In general, females with young were more vigilant than their young or females without young. There were few gender differences in vigilance among animals without young, except that male zebra, wildebeest and waterbuck were more vigilant than females. Animals on the edge of herds devoted more time to vigilance than intermediate or central animals (regardless of sex or age class). Although there was no relationship between group size and vigilance in 1993 when all herds were considered, there was a decrease in vigilance with increasing herd size for impala and kob for herds less than 50. Moveover, herd size and vigilance were negatively correlated in our 1984-1985 study, due partly to differences in group sizes. In 1984-1985 animals were not migrating, and were in relatively small, discrete groups, whereas in 1993 some species were migrating in large herds that stretched for several km. We suggest that in 1993 herd size was above a threshold where increases in group size can lead to further decreases in vigilance. Under these circumstances, location in the herd becomes paramount: Outside animals can be directly exposed to attack. Consequently animals on the edge devote more time to vigilance than central animals.