We investigate the evolutionarily stable seasonal timing of life-history events, such as the start of breeding, when the risk of physical disturbance and advantages of growth or competitive advantage are in conflict. In our model, disturbance is assumed to cause failure of all breeding (or other life history) attempts made before it. For a given probability distribution of the timing of last date of environmental disturbance, individuals starting breeding earlier face a higher risk of being disturbed. On the other hand, the clutch laid earlier has an advantage over those laid later, either having more time to grow or to pre-empt the nest sites, suitable habitat or foraging sites. The evolutionarily stable population is determined to have a single optimal date on which all individuals start breeding (synchronous breeding), provided the benefit of the earlier start of breeding is given purely by the excess time for growth, not by the competitive advantage, and provided offspring produced from many subpopulations, each having a different date of disturbance, are pooled to form a population over which population regulation occurs. In contrast, an ESS population may include a period of breeding over which some individuals start breeding every day (asynchronous breeding), if (i) early breeding is accompanied by competitive advantage within the local population; (ii) population regulation occurs within the local population before being mixed with offspring produced from other local populations; or (iii) disturbance occurs synchronously over the whole population. In short, synchronous breeding should evolve if the spatial scale over which disturbance dates are strongly correlated (denoted by Sdisturbance) is much smaller than the spatial scale of the population in which population regulation occurs (denoted by Sregulation). Asynchronous breeding is expected to evolve if Sdisturbance is similar to or larger than Sregulation.