In this paper, we review briefly the current state of knowledge about sexual differentiation in eutherian mammals, and then describe the situation in detail in two marsupial species: the North American opossum and the tammar wallaby. The conventional explanation for the genesis of all male somatic sexual dimorphisms in mammals is that they are a consequence of the systemic action of testicular hormones. In the absence of testes, the embryo will develop a female phenotype. We present evidence for the tammar wallaby that calls into question the universal applicability of this hormonal theory of mammalian sexual differentiation. We have shown that extensive somatic sexual dimorphisms precede by many days the first morphological evidence of testicular formation, which does not occur until around the third day of pouch life. Male foetuses, and pouch young on the day of birth, already have a well-developed gubernaculum and processus vaginalis, paired scrotal anlagen, and a complete absence of mammary anlagen, whereas female foetuses and newborn pouch young have a poorly developed gubernaculum and processus vaginalis, no scrotal anlagen, and well-developed mammary anlagen. Because it seems unlikely that the male gonad could begin hormone secretion until after the Sertoli and Leydig cells are developed, our results strongly suggest that some sexually dimorphic somatic characteristics develop autonomously, depending on their genotype rather than the hormonal environment to which they are exposed. We have been able to confirm the hormonal independence of the scrotum, pouch and mammary gland by administering testosterone propionate daily by mouth to female pouch young from the day of birth; although the Wolffian duct was hyperstimulated, there was no sign of scrotal development, or pouch or mammary inhibition. When male pouch young were treated with oestradiol benzoate in a similar fashion, there was hyperstimulation of the Müllerian duct and inhibition of testicular migration and development, but no sign of scrotal inhibition or pouch or mammary development. Our results in the tammar wallaby are consistent with the earlier studies on the opossum, whose significance was not appreciated at the time. Further evidence in support of this hormonal independence comes from earlier studies of spontaneously occurring intersexes in several species of marsupial, including the opossum and the tammar wallaby. An XXY individual had intra-abdominal testes and complete masculinization of the male reproductive tract internally, but externally there was a pouch and mammary glands and no scrotum.(ABSTRACT TRUNCATED AT 400 WORDS)