Environmental stochasticity in dispersal areas can explain the "mysterious" disappearance of breeding populations.

We present the results of an individual-based simulation model, showing that increasing the mortality of non-breeding dispersers within settlement areas can lead to the extinction of species and (meta)populations in a subtle way. This is because the areas where dispersers settle are generally unknown or difficult to detect. Consequently, fewer efforts are devoted to the conservation of these sites than to the conservation of breeding territories. Additionally, high mortality rates affecting the floater sector of a population become evident in the breeding sector only after several of years, when it is too difficult or too late to halt the decline. As a result, because most conservation projects on endangered species and populations mainly focus on breeding areas, many current efforts may be wasted in locations other than those in which conservation would be really necessary and effective.