As early as 1887, it was postulated that the mature oocyte possesses all of the elements necessary for embryonic development with the exception of an active division centre, and that the spermatozoon contains such a centre, but lacks the substrate in which to operate. This division centre is called the centrosome. The precise definition of this structure is still a subject for debate. It consists of two centrioles in a perpendicular arrangement and pericentriolar material, and is considered to be responsible for nucleation of microtubules and the formation of the mitotic spindle. There is a paternal pattern of inheritance of the centrosome in humans; thus, human oocytes lack centrioles but the spermatozoa carry two. At gamete fusion the sperm tail is incorporated into the ooplasm, and the centriolar region forms the sperm aster while the sperm head is decondensing; this aster acts to guide the female pronucleus towards the male pronucleus. The centriole duplicates during the pronuclear stage, and at syngamy centrioles are found at opposite poles of the first cleavage. The centrosome has several implications for human infertility. It is possible that immotile or nonprogressively motile spermatozoa may possess centriolar abnormalities or an absence of centrioles. Similarly, antisperm antibodies against centrioles may be responsible for mitotic arrest. One way of solving this problem would be the use of donor centrosomes. To this end, we have assessed the ability of embryos injected with physically separated sperm segments (head only, head and tail separated or isolated tail) to develop normally. Fluorescent in situ hybridization revealed an almost universal mosaicism in these embryos, suggesting that physical disruption of the spermatozoa compromises the ability of the centrosome to function in the zygote. Thus far, centrosome donation with centriole-carrier flagellae obtained by this dissection method does not appear to be feasible.