There have been a number of studies over the past 15–20 years which have suggested that sperm counts in men are on the decline (reviewed in Carlsen et al. 1992). These have been largely ignored, or dismissed on the grounds that they have been subject to selection bias and/or the inclusion of data from men with fertility or testicular problems (e.g. see MacLeod & Wang, 1979). Now comes a new and comprehensive study (Carlsen et al. 1992) which has avoided these 'pitfalls'. These authors have screened every publication between 1938 and 1991 which has reported sperm counts in normal (i.e. unselected) or fertile men - a total of 61 publications with data on 14 947 men. Analysis revealed a highly significant (P6/ml) and 1990 (66 × 106/ml). Moreover, as semen volume also declined