Using positive reinforcement, J. McKinley trained 12 common marmosets (Callithrix jacchus) to provide urine samples on request. The study then exposed the marmosets to mildly stressful, routine husbandry procedures (i.e., capture and weighing). The nonhuman animals spent less time inactive poststressor as opposed to prestressor. L. Bassett collected matched behavioral data from 12 nontrained marmosets who were less accustomed to human interaction. These animals spent significantly more time self-scratching and locomoting as well as less time inactive, poststressor. Collapsed data from the 2 populations showed increased scent marking, poststressor. These results suggest that locomotion, self-scratching, and scent marking are useful, noninvasive behavioral measures of stress and, thus, reduced welfare in the common marmoset. Overall, nontrained animals showed more self-scratching than did their trained counterparts. It was not possible to collect urine from nontrained marmosets. In response to the stressor, however, trained animals showed no significant change in excreted urinary cortisol. These results suggest that training marmosets may allow them to cope better with routine laboratory procedures.