Diagrammatic representation of the hypothetical dynamics of scrapie infection in the herd. Left axis, prevalence of infection. Left arrow, purchase of the original herd. Right axis, number of scrapie cases recorded prior to the cull, which are represented as blue bars; red circle, index scrapie case. Right arrow, time of cull. X-axis, years. Exponential line, hypothetical increase of scrapie prevalence in the herd as a result of mounting pressure of infection, until reaching the 36.0% figure observed in the 200 goats examined (see text). Middle arrow dividing periods A and B is set in 2003 to accommodate for the age split (younger and older than 60 months) used to analyse the effect of age on prevalence. During period A, the prevalence of infection would have been low, mostly involving the most susceptible II₁₄₂ goats; clinical cases would have been very few, unnoticed and undiagnosed. Therefore, most (88.7%) of the 75 scrapie cases detected later (blue bars) were in II₁₄₂ goats (see text). Also, at the time of the cull, the prevalence of infection amongst goats older than 60 months (i.e. born during period A) was significantly higher in II₁₄₂ homozygotes than in methionine carriers (47.4% and 15.1%, respectively, see Tab. IV). Raising pressure of infection during period B would have made young methionine carriers to become susceptible to infection, so that, at the time of the cull, methionine carriers younger than 60 months (i.e. born during period B) showed similar prevalence of infection than II₁₄₂ homozygotes, and significantly higher than methionine carriers born during period A (see Tab. IV). It is assumed that scrapie infection was already present in the herd when originally purchased in 1998, as it is highly unlikely that neither prevalence of infection at the time of culling (early 2008) nor initial detection of clinical scrapie cases in 2005 could have resulted from the incorporation of 100 goatlings to the herd in 2003.