Abstract In a Fear/Defense Test Battery, measuring defensive reactions to a present, approaching and contacting predator, the highest dose of morphine tested (7.5 mg/kg) reliably reduced vocalization to dorsal contact, to vibrissae stimulation, and to an anesthetized conspecific in laboratory-bred wild R. norvegicus. Except for a dose-dependent reduction in flinch/jump reactions to dorsal contact (taps), other defensive behaviors (flight, freezing, etc.) were not reliably altered by morphine treatment (0, 1.0, 2.5, 7.5 mg/kg). Vocalization responses to vabrissae stimulation in wild-trapped R. rattus were reliably increased following naloxone (1.0 and 10.0 mg/kg) administration, lending support for opiate receptor involvement in the mediation of defensive vocalization. In the Anxiety/Defense Test Battery, measuring defensive reactions to situations associated with a predator (cat) or with cat odor, laboratory rats showed no decrease in defensive behavior with morphine (0, 1.0, 5.0 mg/kg). In direct contrast to the above findings, the effects of morphine treatment in this test battery suggested a generalized increase in defensiveness to noncontacting and nonpainful threat stimuli. These effects included a decrease in time spent near the cat compartment, with a complementary increase in time spent at maximum distance, a decrease in transits between these sections, an increase in crouching, and a decrease in grooming and rearing. This pattern of results suggests that morphine may have two opposing effects of defensive behavior, a generalized enhancement, together with a more specific reduction of responses to tactile or painful stimulation. A very widespread pattern of reliable sex or sex × drug effects in the Anxiety/Defense Test Battery was in good agreement with previous report of sex differences in these tests, with females generally more defensive than males. Consonant with previous findings, no reliable sex differences were found with the Fear/Defense Test Battery, although several values approached an acceptable level of statistical significance.