Two cannabinoid receptors, CB1 and CB2, are expressed in mammals, birds, reptiles, and fish. The presence of cannabinoid receptors in invertebrates has been controversial, due to conflicting evidence. We conducted a systematic review of the literature, using expanded search parameters. Evidence presented in the literature varied in validity, ranging from crude in vivo behavioural assays to robust in silico ortholog discovery. No research existed for several clades of invertebrates; we therefore tested for cannabinoid receptors in seven representative species, using tritiated ligand binding assays with [3H]CP55,940 displaced by the CB1-selective antagonist SR141716A. Specific binding of [3H]CP55,940 was found in neural membranes of Ciona intestinalis (Deuterstoma, a positive control), Lumbricusterrestris (Lophotrochozoa), and three ecdysozoans: Peripatoides novae-zealandiae (Onychophora), Jasus edwardi (Crustacea) and Panagrellus redivivus (Nematoda); the potency of displacement by SR141716A was comparable to measurements on rat cerebellum. No specific binding was observed in Actinothoe albocincta (Cnidaria) or Tethya aurantium (Porifera). The phylogenetic distribution of cannabinoid receptors may address taxonomic questions; previous studies suggested that the loss of CB1 was a synapomorphy shared by ecdysozoans. Our discovery of cannabinoid receptors in some nematodes, onychophorans, and crustaceans does not contradict the Ecdysozoa hypothesis, but gives it no support. We hypothesize that cannabinoid receptors evolved in the last common ancestor of bilaterians, with secondary loss occurring in insects and other clades. Conflicting data regarding Cnidarians precludes hypotheses regarding the last common ancestor of eumetazoans. No cannabinoid receptors are expressed in sponges, which probably diverged before the origin of the eumetazoan ancestor.