The ecology of the red-listed butterfly Lopinga achine was studied in partly open woodlands in the province of Östergotland, Sweden. Detailed autoecological research is essential for successful conservation of a species, and the present investigation focused on the initial aspects of such work, namely, choice of host plant, habitat selection, and dispersal ability. The effects of patch area, isolation, and successional stages of studied sites were also examined.

The results indicate that L. achine depends on a single host plant, Carex montana. The females preferred to oviposit near this sedge, and most of the larvae (> 80%) were found on C. montana in the field. Egg-laying females and larvae were restricted to C. montana growing in a narrow zone along the edges of glades. This restriction to forest edges is probably the cause for the dependence of L. achine on a restricted stage of canopy cover. More specifically, no L. achine occurred at sites with less than 60% canopy cover, and population densities decreased sharply with more than 90% cover. Eighty-six percent of the studied sites that were occupied are unmanaged, thus the most important aspect of long-term conservation of L. achine is probably the deterministic changes of its woodland habitat. If the sites remain unmanaged, the system of populations will most likely collapse within 20-40 years. Populations of both L. achine and C. montana increased in size at experimentally managed sites where new glades were created. However, an important prerequisite of successful restoration appears to be the presence of C. montana along the edges of new glades from the start, because the rate of C. montana colonisation was slow.

Most of the populations (50 of 79) were small (< 500 adults; none larger than 4,500) and seemed to show synchronous interannual fluctuation. The probability that a patch would be occupied increased with increasing patch area and decreasing distance to the nearest occupied patch. This was presumably due to different probabilities of extinctions, colonisations, and survival of the inhabiting populations. All but two of the sites with ≥ 3 individuals were within 740 m of the nearest neighbour. Patch size is also a key factor for occurrence: compared to larger patches, small patches are more dependent on neighbour populations.

The majority of the movements were small and within sites, although in many cases the distance to other sites was less than 100 m. Only 56 individuals (4.0% of those recaptured) moved between sites. It seems that habitat patches of L. achine should be less than 700 m from each other to ensure inter-population contact. Fifteen to 20 wellconnected populations have been stated as a lower limit for a viable metapopulation.Based on these criteria, there are two groups of viable populations in the studied area,and these two groups will be given priority in future conservation work.