Basic parameters which are crucial for the survival of human embryonic striatal grafts need to be investigated before initiating clinical trials in Huntington's disease. In order to define the dissection of human striatal-donor tissue which gives rise to the largest amount of striatal neurons after intrastriatal transplantation, we studied the lateral and medial ganglionic eminences of embryonic striatal primordia obtained from human embryos sized 17-30 mm in crown-to-rump length (corresponding to Carnegie stages 18-23). Anatomical landmarks that demarcated the lateral and medial ganglionic eminences from each other were present only in embryos with 20 mm crown-to-rump length or larger. In monolayer cultures, the lateral ganglionic eminence gave rise to a six-fold higher yield of dopamine- and cyclic AMP-regulated phosphoprotein 32-immunoreactive striatal neurons as compared to the medial ganglionic eminence. We also xenografted the lateral and medial ganglionic eminences from five embryos sized 21-30 mm in crown-to-rump length to the ibotenate lesioned striatum of immunosuppressed rats. The grafts were evaluated with respect to general morphology, survival and integration using (immuno-) histochemical stains for acetylcholinesterase/Cresyl Violet, nicotinamide adenine dinucleotide phosphate-diaphorase, dopamine- and cyclic AMP-regulated phosphoprotein-32, tyrosine hydroxylase and calbindin-D28KD. As assessed 9-25 weeks after implantation, 13 out of 16 and 8 out of 13 grafts, in the groups grafted with the medial and lateral ganglionic eminences, respectively, had survived. Previous studies with rat donor tissue have indicated that the functional efficacy of striatal grafts is related to the development of striatal-specific P-zone regions and that these are enriched in transplants derived from the lateral as opposed to the medial ganglionic eminence. Also in the human striatal xenografts of the present study, P-zones appeared more abundant when the donor tissue was derived from the lateral ganglionic eminence. However, the proportion of graft tissue that expressed P-zone properties was always very low (at most 30%) and never approached the 80-90% previously observed in transplants of rat lateral ganglionic eminence. We conclude that the relative yield of striatal neurons in grafts of the human embryonic striatal primordium has to be improved before neural transplantation should be applied in patients with Huntington's disease.