Some reasons why latent period should not be considered constant over the course of a plant disease epidemic

The latent period is a crucial life history trait, particularly for polycyclic plant diseases, because it determines how many complete monocycles could theoretically occur during an epidemic. Many studies have focused on the variation of latent period with pathogen or host genotype, or its plasticity in response to environmental factors. The focus on these aspects is unsurprising, as these factors classically form the apices of the epidemiological triangle. Experiments in controlled conditions are generally used to assess pathogenicity and host susceptibility. Most epidemiological studies consider implicitly that the latent period of a plant pathogen population is constant in time. Once estimated for one or several pairs of host-pathogen genotypes, the value of the parameter ‘latent period’ is usually considered fixed and is often used “as is” in epidemiological models. My thesis here is that the latent period can display non-negligible variability over the course of a disease epidemic, and that this variability has multiple sources, some of which have complex, antagonistic impacts. I develop arguments for four sources of variation challenging the implicit assumption that the latent period remains constant: daily fluctuations in leaf temperature, nature of inoculum, host stage or age of host tissues, intrapopulation competition and selection for aggressiveness traits. I focus on the wheat fungal pathogen Septoria tritici blotch (Zymoseptoria tritici), making use of empirical datasets collected during my own research projects and a targeted literature review. Finally I discuss in which cases certain sources of these variations should be accounted for into epidemiological models.