The sensitivity of local populations to year-to-year differences in rainfall combined with the enormous variation in timing and amount of rainfall this area receives suggests that they are highly susceptible to extinction, as does the observation of several local extinctions and re-colonizations in the past 30 years. The best explanation for the persistence of many of the small, isolated populations is that they form part of an extended metapopulation linked by recurrent extinctions and recolonizations. It is interesting to wonder whether these metapopulation dynamics have always been characteristic of the species or if they date only from the introduction of non-native grasses and the reduction of native grasslands to their current patchwork (ca. 1800).
In the Morgan Hill population (Santa Clara County; Figure 1), one population is far larger than any of the remaining ones. It apparently acts as a reservoir from which smaller patches are recolonized. Interestingly, butterflies were found on eight small patches of native grassland within 5 km of this apparent source, but none were found on more distant patches that appear to be suitable habitat. Immigration rate estimates were derived from a model assuming that
This analysis suggested that satellite populations were unlikely to be maintained beyond a radius of 7 km, because catastrophic extinction occurs at a higher rate than re-colonization beyond this distance.
Stepping-stone migration among satellite populations contributes relatively little to population persistence because