Amphetamine induces a pronounced rotation directed ipsilateral to the lesion and lasting about 2 h in rats bearing a unilateral lesion of the nigrostriatal dopaminergic pathway. Implantation of embryonic dopaminergic neurons into the lesioned striatum leads to a compensation of this rotation. However, graft-bearing animals display a strong biphasic contralateral rotation, lasting up to 5 h. To try to ascertain the mechanisms of this anomalous rotation, two separate experiments were performed. First, we tested whether the contralateral rotation presented by the grafted animals could be correlated to the persistence of the lesion-induced decoupling of striatal D 1 and D 2 receptors. Lesioned and grafted animals were submitted to a series of four amphetamine (5 mg/kg, i.p.) rotation tests. Prececling each test animals received, in a randomized order, one of four of the following treatments: physiological saline, a D 1 receptor blocker (SCH-23390, 0.1 mg/kg, s.c.), a D 2 receptor blocker (raclopride, 2.5mg/kg, i.p.) or the combination of the D 1 and D 2 antagonists. The ipsilateral rotation observed in the lesioned animals was abolished by the separate blockade of both classes of dopamine receptor as well as by their combined blockade. Grafted animals could be separated into two subgroups, based on the effect of the antagonists during the first 2 h of amphetamine-induced rotation. In one subgroup, antagonists had the same effect on the amphetamine-induced contralateral rotation as they did on the ipsilateral rotation displayed by lesioned animals. In this group, D 1 and D 2 receptors were therefore recoupled by the implant in the lesioned striatum. In the other subgroup, the contralateral rotation could be antagonized only by the combined D 1 and D 2 blockade, while the separate blockade of D 1 or D 2 receptors did not decrease or even increased the amphetamine-induced rotation. This indicates that in this group the lesion-induced decoupling of D 1 and D 2 receptors persisted. Nevertheless, the characteristics of the amphetamine-induced rotation (magnitude, duration) were the same in the two subgroups. Likewise, hypersensitivities of both D 1 and D 2 receptors were completely abolished by the graft in both subgroups. From this experiment it is concluded that the amphetamine-induced rotation observed in grafted animals is not correlated with the state of coupling of striatal D 1 or D 2 receptors. In a second experiment, dopamine release was monitored by microdialysis in the graft-bearing and the contralateral normal striatum of awake, behaving animals following the administration of amphetamine to test whether the observed rotation could be explained by a higher than normal dopamine release from the implanted dopaminergic neurons. Grafting restored both basal (about 50% of normal values) and amphetamine-induced (between 60 and 80% of normal values) dopamine release in the graft-bearing striatum. However, amphetamine-induced dopamine release on the graft-bearing side never exceeded that measured in the contralateral striatum, even though the animals displayed, at the time of collection of the fraction, the characteristic contralateral rotation. From these experiments it is concluded that the amphetamine-induced rotation observed in grafted animals is not related to some abnormal presynaptic mechanism or to a modification of the state or coupling of postsynaptic D 1 or D 2 receptors. The cause of this anomalous rotation should therefore reside in some abnormal postsynaptic reactivity to an otherwise normal dopaminergic stimulation originating in the grafted neurons.