Catastrophic extinction of population sources in a butterfly metapopulation

Increasing emphasis is being placed on the large-scale and long-term dynamics of populations. A butterfly (Euphydryas editha) metapopulation that was naturally restricted to rocky outcrops in an area of coniferous forest suffered two major perturbations in 30 yr. First, humans clear-cut patches of forest in about 1967. The butterfly colonized the clear-cuts, in which it began to oviposit on a novel host plant, Collinsia torreyi. Breeding success was high, and the clear-cut habitat supported a population source; in the mid-1980s, there was net movement of adults from clear-cut to outcrop, and butterfly densities were elevated on outcrops close to population sources. The second major perturbation came in 1992, when a severe summer frost killed C. torreyi plants (outcrop hosts were not damaged). As a result, E. editha larvae starved in clear-cuts, and the source populations became extinct. This occurrence allowed us to test several predictions of source-sink theory. As predicted, densities declined on outcrops, and particularly on those close to former sources; by 1993, densities on outcrops were no longer correlated with isolation from former sources. True population sinks are predicted to become extinct in the absence of immigration, but this prediction was not observed. Outcrops were not true sinks because E. editha occupied this habitat before the creation of sources, survived the frost on outcrops, and persisted on outcrops in areas where they were isolated from sources. Outcrops were pseudosinks; breeding success was poor at immigration-enhanced densities, but outcrop populations did not rely on immigration to persist. The implications for conservation biology are that populations with high mean density and that act as sources are not necessarily more stable than lower-density or pseudosink populations. This may be especially true of systems subject to recent human disturbance, as in the present case.