This paper briefly reviews the literature on testicular descent and presents new observations from the authors' laboratory which suggest new ways of looking at old problems. There is now good evidence that testicular descent occurs in two morphologically and hormonally distinct phases. Relative 'transabdominal migration' of the testis compared with the ovary occurs at 10-15 weeks of gestation in the human and 'inguinoscrotal' migration occurs at 26-35 weeks of gestation. We have proposed previously that the first phase is controlled by Müllerian inhibiting substance although this remains controversial. The second phase is androgen dependent and is possibly mediated indirectly through the release from the genitofemoral nerve (GFN) of the neuropeptide calcitonin gene-related peptide (CGRP). Recently we have used three different rodent models of undescended testis to determine the involvement of the GNF and/or CGRP. The testicular feminization mouse with complete androgen resistance and the rat exposed prenatally to the antiandrogen flutamide have a deficiency of CGRP in the GFN. In contrast, the mutant trans-scrotal rat which has normal androgen levels has an excess of CGRP in the GFN. All cryptorchidism models, despite their different primary cause, have in common an abnormality of the GNF and/or CGRP which is consistent with the hypothesis that normal testicular descent in the rodent may be mediated by the GFN.