Five nursery‐reared rhesus (Macaca mulatta) infants were studied on 8 consecutive days over a period of 2 weeks when they ranged from 4 to 5.5 months. Saliva cortisol samples were obtained by giving each animal, while in its individual home cage, a 6‐in. cotton dental roll pretreated with sugared fruit‐drink crystals and dried. Subjects were allowed to mouth the cotton roll for 10 min during each collection period. Saliva was then expressed from the cotton and analyzed by radioimmune assay. Samples were collected at 0830, 1100, 1400, and 1630 hr on each of the 8 days. On four of the days, 2‐hr peer socialization sessions were imposed between 1200 and 1400 hr. Significant time‐of‐day effects were obtained. Values tended to be low at 0830 hr, rising significantly to peak levels at 1100 hr, and then declining over the 1400‐hr and 1630‐hr sampling periods. All 4 subjects with sufficient samples had higher average cortisol concentrations when the 1400‐hr sampling followed peer‐socialization sessions, compared to no‐socialization days. This difference was only marginally significant by paired t test, however. These results (a) support the feasibility of using noninvasive salivary sampling procedures with infant monkeys, and (b) indicate that salivary cortisol measures are sensitive to daily rhythms. Unexpectedly, the results also raise the possibility that nursery‐reared infant rhesus macaques may be phase‐delayed in their HPA rhythm, with the morning peak occurring several hours after rather than in the hour before lights‐on and the morning feeding. © 1995 John Wiley & Sons, Inc.