This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues.Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. The period between sowing and the point of time, when 50% of the leaf area was necrotized due to colonization by S. tritici (T 50 ) was estimated for each year, site, cultivar and leaf layer by non-linear regression. T 50 values followed a bimodal distribution with one maximum at 245 days after sowing (DAS; early epidemics) and one maximum at 270 DAS (late epidemics). Early epidemics were preceded by almost constant daily average temperatures of 13.2 AE 0.8 C between 181 and 210 DAS. Late epidemics were preceded by an approximately linear increase in temperature from 8.7 AE 0.9 to 12.1 AE 0.9 C during the same period of time. Based on these differences, it seems possible to predict whether an early or a late epidemic can be expected at least 35 days before the epidemic outbreak. Temperature sums calculated with a base temperature of 6.6 C starting at sowing and ending when T 50 was reached were not significantly different between early and late epidemics (P ¼ 0.73) and averaged 1721 AE 49 days. Fungicide applications, which resulted into a delay of the epidemic development similar to the difference between early and late epidemics, resulted in a yield increase between 11.7 and 12.6%.