The objective of the studies was to evaluate infertility according to sperm count shifts. The distribution of the sperm count of 1024 Danish men (median 56 million/ml) served as reference. The data were transformed with multiplicative or additive models to create alternative distributions with median sperm count values changed by 25-100%. Sperm-count-specific fecundabilities were provided from a follow-up of first-pregnancy planners in a Danish population. The estimated average fecundability of the 1024 Danish men was 16.9% [95% confidence interval (95% CI) 16.7-17.2], and the proportion of cohabiting men with spouses pregnant within 1 year was 86.0% (95% CI 84.1-87.8). Simulations of alternative sperm count distributions indicated that the relationship between sperm count shift and fertility strongly depends on the median level of the sperm count at onset and the type of shift, a dramatic decline from a high level in a multiplicative model indicating a marginal change and a minor decline from a low level in an additive model representing a strong decrease in fertility. In some cases sperm count, therefore, may be an early warning of changes in fertility.