There are contradictory results concerning the effects of systemic injections of cannabinoid agonists in anxiety-induced behavioral changes. Direct drug administration into brain structures related to defensive responses could help to clarify the role of cannabinoids in these changes. Activation of cannabinoid CB(1) receptors in the dorsolateral periaqueductal gray induces anxiolytic-like effects in the elevated plus maze. The aim of this work was to verify if facilitation of endocannabinoid-mediated neurotransmission in this region would also produce anxiolytic-like effects in another model of anxiety, the Vogel conflict test. Male Wistar rats (n=5-9/group) with cannulae aimed at the dorsolateral periaqueductal gray were water deprived for 24 h and pre-exposed to the apparatus where they were allowed to drink for 3 min. After another 24 h-period of water deprivation, they received the microinjections and, 10 min later, were placed into the experimental box. In this box an electrical shock (0.5 mA, 2 s) was delivered in the spout of a drinking bottle at every twenty licks. The animals received a first microinjection of vehicle (0.2 microl) or AM251 (a cannabinoid CB(1) receptor antagonist; 100 pmol) followed, 5 min later, by a second microinjection of vehicle, anandamide (an endocannabinoid, 5 pmol), AM404 (an inhibitor of anandamide uptake, 50 pmol) or URB597 (an inhibitor of Fatty Acid Amide Hydrolase, 0.01 or 0.1 nmol). Anandamide, AM404 and URB597 (0.01 nmol) increased the total number of punished licks. These effects were prevented by AM251. The results give further support to the proposal that facilitation of CB(1) receptor-mediated endocannabinoid neurotransmission in the dorsolateral periaqueductal gray modulates defensive responses.