Tallgrass prairie butterfly surveys in recent decades in four states in the USA indicate numerous declines of prairie-specialist butterflies including Speyeria idalia, Oarisma poweshiek, Atrytone arogos, Hesperia dacotae, and H. ottoe in fire-managed preserves, including large high-quality ones. These results replicate previous findings, indicating that upon initiation of conservation action, both cessation of prior management and inception of new management affect specialists negatively and that butterfly declines can be as great on reserves as nonreserves. Results at Wisconsin sites with species-specific management protocols, including permanent non-fire refugia, were more favorable for the specialists (S. idalia, Lycaeides melissa samuelis) the protocols were specifically designed to benefit. Butterfly declines after preservation will likely continue unless the conservation approach changes to include consideration of individual species' required resources and management tolerances. The ecosystem approach assumes that habitat specialists are co-evolved with processes such as fires assumed to maintain those ecosystems. Data presented here indicate that tallgrass prairie specialist butterflies are not co-evolved with current fire regimes. An alternate perspective views ecological processes as resetting vegetation to current climate and landscape conditions. Over geologic time, relict vegetation associations persist as outliers until an event resets them. In modern times, human disturbances (especially soil-exposing ones) can reset sites to favour the more generalist species (plants and butterflies) found in the prevailing, human-degraded landscape.

From the early 1990s through 2005, we conducted butterfly transect surveys annually at the same sites in three regions of Wisconsin. We compared specialist butterfly population indices at three sites where a permanent non-fire refugium (a unit kept unburned through cycles of rotational fire elsewhere in the site) was established during this study to indices at comparison sites (which had consistent management throughout this study) in the same region. At Crex Meadows (12,180 ha), all significant changes in specialist numbers skewed toward relatively higher abundance in the refugium (14 ha, last burned in 1988) during the later period (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005) versus 10 comparison subsites (fire-managed in both periods) than expected from observations in the earlier period . The same pattern occurred in the refugium (4 ha) at Bauer-Brockway (125 ha), after the rest of that unit (9 ha) had its first management fire. Population trends were positive in these refugia, while the comparison sites usually had less favorable trends, or otherwise had similar trends. At Muralt Bluff (25 ha), regal fritillaries (Speyeria idalia) were more concentrated in the refugium (3 ha, last burned in 1991) during the earlier period but were more abundant in both periods in the refugium than the other units there (fire-managed in both periods). In the earlier period at Muralt Bluff, this species significantly declined, the opposite trend of comparison sites (which always had never-burned refugia), but significantly increased in the later period, similar to comparison sites. The refugium did not benefit Ottoe skipper (Hesperia ottoe), which declined significantly in the earlier period and was not recorded in the later period. Formerly burned units began functioning as refugia only >6-8 years after last fire and continued to increase in benefit for years after that. In fire-managed and fireprone sites, we recommend the creation of permanent non-fire refugia for Lepidoptera conservation, placed where the most specialists will benefit and managed unintensively (e.g. mowing) if needed.

SummaryThree species of declining grassland songbirds were counted in June transects at 42 south-western Missouri (U.S.A.) prairie preserves managed three different ways: biennial July haying, spring burning on a 1-4 year cycle, and sites both burned and hayed every 2-4 years. Hayed prairies had twice as many Henslow's Sparrows Ammodramus henslowi as burned prairies and 2.5 times as many as sites which had been burned and hayed. Hayed prairies had 59% more Grasshopper Sparrows A. savannarum and 14% more Dickcissels Spiza americana than burned prairies. Fired and hayed sites had 34% higher detection rates of Grasshopper Sparrows than sites only burned and similar numbers of Henslow's Sparrows and Dickcissels as burned sites. The three species summed were 60% more abundant in hayed as in fired or fired and hayed prairies. Henslow's Sparrows increased and Grasshopper Sparrows decreased with time since the last management. Grasshopper Sparrows peaked one year earlier after haying than after burning. Dickcissels did not vary by time since last management. Henslow's Sparrows and Dickcissels were significantly more abundant in larger prairies. Preserve management for these three prairie birds should emphasize biennial to triennial mid-summer haying, instead of burning.

During 1993-1996, two teams (Schlicht, Swengels) surveyed the same Minnesota prairies, but without any coordination of sites, routes, methods, dates, and results between teams. In 27 instances, both teams surveyed the same site in the same year between 30 June and 18 July. For the 18 most frequently recorded species, abundance indices (individuals/h per site) significantly covaried between teams for 11 (61%) species, including 2/3 prairie specialists tested. No species significantly correlated negatively, 17/18 species had positive correlations, and the preponderance of positive correlations was significant. Swengel indices per hour (two surveyors; unlimited-width transect) averaged 2.42 times Schlicht indices (one surveyor; fixed-width transect). These results demonstrate that transect surveys by different teams at the same sites but not the same routes produce similar rankings of species abundance among sites. This approach to population monitoring (transect surveys during the season that covers the most specialist species at once, not necessarily with fixed routes but recording all species seen) might also be appropriate in other regions with high habitat loss and low human population density. Abundance indices from surveys by seven teams spanning 1979-2005 were calculated for evaluating population trends. For the five analyzable specialist species, 25/30 population trend tests of a species at a site had a negative direction, a highly significant skewing (P \ 0.0001). By contrast, five ''common'' (most frequently recorded non-specialist) species had an even distribution of negative and positive trends. While adjacent sites had similarly timed decline thresholds (last year when a higher rate or any individual was recorded vs. first year when all subsequent indices were lower or zero) within species, these thresholds were not synchronized among sites in different counties. All sites analyzed in this study were preserves managed primarily with fire. While the ecosystem (or vegetative) approach to reserve selection has been validated in other studies to be effective at capturing populations of associated specialist butterflies, butterfly declines after reserve designation will likely continue unless the ecosystem approach to reserve management includes specific consideration of individual butterfly species' required resources and management tolerances.

No abstract