Table 2. Specification of the management options tested.

Management option	Management action		Management cost	Implementation in the model	Time lag of effect
Habitat enlargement	From adjoining degraded wet heathland/forest: sod-cutting, liming, removal of trees and prevention of drainage (filling ditches)		5,000 €/ha 1	Existing patches were enlarged to 0.5 ha when smaller (viable populations are sustained by >0.5 ha for M. alcon 2). When more than 4 ha (20,000 €) would be needed to enlarge all existing patches, preference was given to occupied patches. Final habitat quality was set to average.	Habitat quality is increasing linearly over 20 years before the target K is achieved 2
Habitat quality improvement	Small-scale sod-cutting + liming, sowing of gentian seed, improving hydrology, adjusted grazing or mowing, controlling tree establishment		2,500 €/ha for 1 quality category up, irrespectively of the current habitat quality category 1	Patches with habitat quality lower than a certain threshold category (Ballooërveld and Drents-Friese Wold: excellent; Delleburen: fair; Dwingeloo: average), defined according to the initial habitat quality within each network and the 20,000 € budget limit.	A linear increase in K during the next 5 (one category up), 7 (2 categories up), 10 (3 categories up), or 12 years (4 categories up) 2
Stepping stones creation	From agricultural land: topsoil removal of 50 cm and introduction of cut heather with gentian seed.	From degraded wet heathland/forest: sod-cutting, liming, removal of trees and prevention of drainage (filling ditches)	15,000 €/ha from agricultural land 1 and 5,000 €/ha from degraded wet heathland/forest 1	A series of 0.5 ha average habitat quality patches were created in the landscape matrix within each network. For each new patch to be created, the current land use at its exact position was determined. The number of stepping stones (4–6) differed between the networks as a function of the cost difference between the restoration of patches from agricultural (15,000 €/ha) and degraded heathland/forest areas (5,000 €/ha). Stepping stones were placed in poorly connected locations (inter-patch distance >2 km maximum dispersal distance 3) in order to connect an occupied patch with a vacant patch of sufficient carrying capacity (wherever possible). The distance to each of the two patches the stepping stone aimed to connect was less than 1.5 km (0.25–1.3 km with a similar c. 0.7 km average for the four networks).	A linear increase in K during the next 20 years, preceded (agricultural land only) by a 10 yr lag 2,4
Reintroduction	A total of 14 captive-reared individuals were reintroduced into vacant patches of each network during 8 years		2,300 € for 14 butterflies per network per year	A minimum of four individuals were introduced in a given patch, with higher priority given to the patches presenting a higher carrying capacity.

1 Rob van der Burg (Bosgroep-Zuid) and René Gerats (Stichting Het Limburgs Landschap, personal communication); www.groenblauwediensten.nl.

2 WallisDeVries 2004 [18].

3 Maes et al. 2004 [15].

4 WallisDeVries unpublished data.