The yield of surviving dopamine cells in nigral grafts is typically low. It is unclear whether the dopamine neurons that do survive are postmitotic at the time of implantation, or are precursor cells that differentiate into dopamine neurons following transplantation in the host brain. We have therefore compared the survival of dopamine neurons in grafts that have been labelled with BrdU at different times prior to or following implantation in order to identify those cells that undergo final cell division at each stage of the procedure. Seven groups of rats were prepared with unilateral nigrostriatal lesions. Three groups received nigral grafts derived from E14 embryos labelled with BrdU in utero on either E12, E13 or E14 days of embryonic age (the E14 injection made 2 h prior to preparation of the graft cell suspension). Three further groups received nigral grafts from untreated E14 embryos, and then dividing cells within the grafts were labelled by injection of BrdU into the host lateral ventricle, 2 h, 1 day or 2 days after implantation (equivalent to E14, E15 and E16 days of embryonic age). The control group received standard (unlabelled) E14 grafts. Five weeks after the transplantation surgery, the host brains were processed using double immunohistochemical techniques to detect tyrosine hydroxylase (TH)-positive neurons which had incorporated BrdU. In the grafts labelled with BrdU prior to implantation, there was an increasing proportion of double-labelled cells (out of the total TH-positive cells surviving in the grafts) with birth dates on E12, E13 and E14 (1%, 12% and 10% per day, respectively). By contrast, grafts labelled following implantation, although containing many dividing neurons, had very few of these BrdU-labelled cells expressing a dopaminergic phenotype; < 1% surviving TH-positive cells were double-labelled from the 2 h post-transplant injection, and < 0.1% from each subsequent injection. This suggests not only that the great majority of TH-positive neurons in nigral grafts were already differentiated at the time of implantation, but also that transplantation of E14 ventral mesencephalic tissue either kills dopaminergic precursors or (more likely in our opinion) prevents their differentiation into a dopaminergic phenotype. Precursor cells that would differentiate into dopaminergic neurons beyond E14 if left in situ in the intact ventral mesencephalon do not readily differentiate into mature dopamine neurons following transplantation. If we are to enhance yields of functional dopamine-rich transplants, then we must identify strategies both to protect predifferentiated dopamine neurons in the grafts and to promote differentiation of a dopaminergic phenotype in precursor cells that continue to divide within the grafts following transplantation into an adult host environment.