The butterfly fauna of Wisconsin bogs: lessons for conservation

Nekola (1998) reported significantly different bog butterfly faunas in the three different bog vegetation types. Even with many more years of surveys, our results on which species occurred in which bog types are remarkably similar to Nekola’s (1998) (Table 4). The minor differences in fauna between Nekola (1998) and us are easily attributable to species accumulation as a function of survey effort (Rosenzweig 1992); more species ought to be found with more visits in more years. European bog Lepidoptera communities exhibit very large differences between different bog structures, different parts of bogs, and between opener bogs and immediately adjacent wet forest (Spitzer et al. 1999; Dapkus 2004a, 2004b).

Nekola (1998) reported significantly fewer bog butterfly species in smaller bogs (muskegs and kettleholes only), but no difference in species richness among the three bog types when controlling for site size. We found that northern Wisconsin bogs were not depauperate in specialists compared to large barrens and heaths in the same region (cf. Table 5, 6). Furthermore, a number of bog specialists frequently occurred in numerous examples of bogs, including all three types (Table 7).

As reported for tyrphobiontic Lepidoptera elsewhere (Väisänen 1992; Spitzer et al. 1999; Dapkus 2004a), specialist species here comprised a small proportion (10%) of all species recorded in bogs (Table 2), similar to the proportion of specialists in three tallgrass prairie subregions (9–16%) and Wisconsin barrens (11%) (Swengel 1998a). However, specialists and affiliates (tyrphophiles) are often the most abundant species in bogs (Väisänen 1992; Spitzer et al. 1999; Dapkus 2004a). In our study, four of the eight specialists were among the six most abundant butterfly species in bogs, out of 77 species recorded (Table 2). Six of the seven most abundant species were bog affiliate and specialist butterflies treated in Nekola (1998) as peatland-obligate species (cf. Table 4).

Specialists accounted for nearly half the total individuals observed in bogs (Table 3). By contrast, only 6% of individuals were specialists in the most fragmented tallgrass prairie subregion, and only 11% in the subregion with the largest patches, while the subregion with both relatively large patches and the most favorable management had 56% specialist individuals (but the seasonal sampling period was the narrowest here, timed for peak specialist numbers) (Swengel and Swengel 2001). Wisconsin barrens (also less fragmented) had 46% specialists (Swengel and Swengel 2001). High fragmentation in a relatively natural landscape due to long-term climatic variation (northern Wisconsin bogs) has more favorable outcomes for specialist butterfly abundance than anthropogenically highly fragmented vegetation (tallgrass prairie). This appears attributable to the high long-term stability of bog vegetation (when relatively undegraded by human activity) (see “Introduction”) that is highly resistant to infiltration by vegetation in the surrounding landscape.

The use of non-native nectar in lowland roadsides by the summer specialists (Table 8) represents a very limited opportunism. The three summer species frequented adjacent lowland roadsides but virtually no individuals of any specialists occurred in adjacent uplands (Table 2). Thus, these species did not in any numbers follow this nectar availability into uplands, where these non-native (as well as native) nectar plants also occur widely.

The vegetative approach to defining habitats may be called the “ecosystem” approach but it’s typically focused on floristic composition (Wisconsin Department of Natural Resources 1995, Packard and Mutel 1997; Panzer and Schwartz 1998). This approach has value: some bog specialist butterflies have remarkable frequency of occurrence in northern Wisconsin bogs (Table 9). This faunistic similarity of specialists across these bogs may be particularly pronounced due to the long-term stability typical of this vegetation and remarkably pristine condition of these sites (see “Introduction”). “Characteristic” butterflies are frequently identified for “zones” or “biomes” (e.g., Layberry et al. 1998, pp 9–11); on that large scale, these are typically “matrix” butterflies of a general vegetation type. But even in highly destroyed and fragmented tallgrass prairie, characteristic specialists (if in range) occurred in many examples of that vegetation (Speyeria idalia in Missouri and Minnesota, Oarisma poweshiek in Minnesota) (Swengel 1998b; Swengel and Swengel 1999a, 1999b). Thus, vegetative classifications are efficacious at grouping insects by their floristic associations (e.g., Panzer and Schwartz 1998; Shuey 2005).

Confounding the fit of butterflies to vegetative classifications is the single vegetative label typically assigned to a particular patch. Numerous non-specialist (non-tyrphobiontic) species associated with other types of vegetations occur in bogs (Table 2). In other words, a northern Wisconsin bog is also a heath that’s wet (Colias interior), a peaty sedge meadow (Satyrodes eurydice), a particularly damp grassland (Coenonympha tullia) or meadow (Boloria selene), and a forest however scraggly (Erynnis icelus, Speyeria atlantis). Even many tyrphobiontic species occur there not because it’s a bog (wetland) but because it’s adequately analogous climatically and vegetatively to taiga or tundra (Spitzer and Danks 2006). At the central Wisconsin pine barrens (Bauer-Brockway) richest in specialist butterfly species in our study (Swengel 1998a), we can record (in season) in one small location specialists of grassland (Hesperia metea, Atrytonopsis hianna, H. leonardus) and savanna (Callophrys irus, C. henrici) as well as forest-specific species (Megisto cymela, Enodia anthedon) (Swengel 2009). This is easily explained by the resource-based approach to defining habitat (Dennis and Eales 1997; Thomas et al. 2001; Dennis et al. 2007; Dennis 2010): each species finding the conditions and resources required. Conservation management decisions can foster or reduce this layering of vegetation types and associated insect diversity on top of each other (Kirby 1992). For example, tree-cutting can maintain a savanna (instead of grassland or forest) at the scale of a site but result in primarily grassland and forest subsites within that site (dissociating the grassland and forest butterflies, and leaving very little partially shaded vegetation for savanna butterflies) or maintain the mix throughout the site at the microsite scale. It is also more difficult to assert that species are co-evolved with a specific regime of processes particular or unique to each ecosystem (e.g., Wisconsin Department of Natural Resources 1995; Packard and Mutel 1997; Panzer 2002) when a site can have more than one ecosystem layered right on top of each other (Kirby 1992).

Aiming for “ecosystems” in conservation management and restoration (Wisconsin Department of Natural Resources 1995; Packard and Mutel 1997; Panzer 2002) is aiming for a native but general vegetation type. This can lead to a more generalist fauna with the loss of specialists (e.g., Kirby 1992; Swengel 1996; Longcore et al. 2000; Nekola 2002). The primary method for prairie conservation management is burning, and this shift can be explained away (sites too small, too degraded) and blamed on the specific method of fire (fires too big, too frequent, and taking away from investing in other kinds of management). But in addition to those factors, an even more fundamental issue is aiming for the average and general ecosystem. Although native, this can lead to average and general butterflies. Bog butterflies reliably live only in sites persistently far outside the landscape average, even in a relatively natural northern Wisconsin context. An alternative approach to both site selection and management embraces site and species individuality by targeting specialists first. For example, by picking spots for the most specialized and rare birds first, then working up from there, all bird species were quickly captured in the fewest sites, compared to other methods of site selection (Williams et al. 1996). By logic, these sites should be conserved for their uniqueness, not be made more typical or average, even if also natural.

To be sure, bogs are particularly long-lived stable vegetation. Other vegetations are naturally more dynamic, so that conservationists aim to conserve and restore that dynamism. But pockets of remarkable stability are natural in other vegetations as well. Brown (1997) described paleo-environments in the tropics particularly speciose in the most conservative insect species, where surprisingly small perturbations of pristine vegetation might have permanent negative effects on those species. In a study of Canadian boreal forest (Gandhi et al. 2001), fire skips (“residuals”) within the perimeter of the most recent wildfire contained older trees than in the unburned forest surrounding the most recent wildfire. The trees in the skips were on average 180 years old (maximum over 300 years old), while the surrounding mature forest unburned in the last fire was only about 72 years old. These fire skips were reservoirs for forest beetles, and the only place where a glacial relict beetle was found. These stably consistent pockets occur in other vegetation types, which also need to be conserved for insects there.

Gandhi et al. (2003) compared the beetle fauna in the fire skips to “residuals” (or “skips”) within a recent clearcut. The clearcut residuals weren’t selected for being “old-growth” and unsurprisingly, the clearcut skips didn’t have the fauna of the wildfire skips. These results do however suggest that clearcut skips could be made more effective for conservation by targeting old-growth (not merely mature) forest. Insects have been proposed as indicators of many things (as reviewed in McGeoch 2007), but a particularly useful property of species groups with adequate knowledge of their ecology would be indication of these outlier paleo-environments not otherwise as easily discerned by plant composition and structure alone. A corollary to Haldane’s possibly apocryphal quip about the creator’s “inordinate fondness for beetles” (as repeated in Ashworth 2001) is an inordinate fondness for specialists (and thus the stability most likely to favor persistence of such faunas), at least given proclivities for landscape dynamism both in the non-conserved modern landscape and in ecological conservation management.

More continuous and unintensive managements (e.g., light grazing) and consistent managements, even if somewhat more intensive (e.g., biennial haying), are more favorable for specialist insects than either intensive or inconsistent managements (Kirby 1992). In rural Sweden, historical land use over the last two centuries was more effective than current land use at explaining which plant species currently lived in the grasslands (Gustavsson et al. 2007). While long-term grazing produced the most favorable floristic results currently, a consistent use of haying throughout the entire period was more favorable than switching from haying to grazing, even decades ago. Thus, conservation management needs to be retrospective to before preservation in embracing site stability (Whitehouse 2006), rather than only forward-looking after preservation and restoration begin. Attempting to turn the clock back to before anthropogenic degradation (or before a switch to less favorable management such as haying in Sweden) can do more harm than embracing and managing to maintain the semi-natural condition of the site now (Kirby 1992). Relatively more stable site histories (e.g., long-term occupancy and cutting by beaver Castor canadensis) also occur for patches occupied by species such as Gillett’s checkerspot (Euphydryas gilletti) well known to inhabit patches generated in a dramatic cyclical way (stand-replacing fire) (Williams 1988). In conserved semi-natural vegetations, more consistent management (grazing) may produce higher relative numbers of localized insects than more dramatic, rotational management (Kirby 1992; Thomas and Harrison 1992).

It is axiomatic that increased plant diversity, especially native, increases insect biodiversity, from gardens to nature reserves (e.g., Panzer and Schwartz 1998; Burghardt et al. 2009). The mechanism for this is not, however, simply plants causing insects, a concept which results in a focus on floristic management and restoration (ecosystems) as the best way to conserve the insect species most associated with those plants. Instead, appropriate, consistent long-term landscape conditions cause both the plants and the associated insects. Just as plants and insects got “sunk and dunked” together in temperate-zone bogs as relicts due to climatic oscillations (Dapkus 2004a; Spitzer and Danks 2006; Whitehouse et al. 2008), so too only insects finding consistent resources in the surrounding landscape exist to benefit when native plants are restored to a garden or reserve. Whatever shortfalls of such resource consistency determine what insects do not benefit from such plantings. A focus on plants can lead to restoration that destroys the continuity of required resources in the process, and loses the associated insects (Kirby 1992), usually the ones most restricted to that site in the first place (such as described in Whitehouse et al. 2008). An alternate approach focuses on what’s “right” about those plants and conditions now (what’s been adequately consistent, however minimally, in resources to maintain such insect faunas), and maintaining that consistency, even if there are “wrong” things too.