Southeastern Naturalist
317
M.M. Jackson, M.G. Turner, and S.M. Pearson
22001144 SOUTHEASTERN NATURALIST 1V3o(2l.) :1331,7 N–3o3. 62
Logging Legacies Affect Insect Pollinator Communities in
Southern Appalachian Forests
Michelle M. Jackson1,*, Monica G. Turner2, and Scott M. Pearson3
Abstract - Many temperate deciduous forests are recovering from past logging, but the
effects of logging legacies and environmental gradients on forest insect pollinators have
not been well studied. In this study, we asked how pollinator abundance and community
composition varied with distance from logging roads and elevation in old (logged >90 years
ago) and young (logged 20–40 years ago) southern Appalachian forests. Insect pollinators
were sampled at 15 previously logged sites along an elevation gradient at 5 distances from
logging roads during summer 2010 and spring 2011 using pan traps. In summer, many pollinator
groups were more abundant in younger forests and closer to logging roads, likely due
in part to more light availability and a greater abundance of floral resources near roads. Total
bee abundance was greater near logging roads, but only in younger forests, suggesting that
the role of roads in providing nectar and other resources may diminish as forests mature. In
spring, many pollinator families were less abundant at mid-distances (2–10 m) from roads
compared to road edges (0 m), but abundances were generally the same at 100 m from the
road as at road edges. Two important bee families, Apidae and Andrenidae, were strongly
associated with high elevations in spring. Our results suggest that logging legacies may
provide supplemental resources such as food and nesting sites to insect pollinators during
the summer months especially, with the effects of roads often extending at least 100 m into
young forests.
Introduction
The importance of insect pollinators for maintaining plant diversity in forests
is well known. At least 67% of flowering plants depend on insects for pollination
(Kearns and Inouye 1997, Ollerton et al. 2011), including a majority of spring
wildflowers in deciduous forests of eastern North America (Motten 1986). There
is evidence for a global pollinator-decline (Biesmeijer 2006, Kearns et al. 1998)
likely related to habitat alteration (Aizen et al. 2003, Cane and Tepedino 2001),
climate change (Memmott et al. 2007), or some combination of the two (Kremen
and Ricketts 2000, Warren et al. 2001), along with other stressors such as disease
and pollutants (Potts et al. 2010). A decline in pollination services could ultimately
lead to extinction of some plant species, especially those that lack compensatory
traits to deal with reproductive failure (Bond 1994).
Currently, most studies of pollinator diversity and abundance focus either on
tropical regions or agricultural fields. It is true that pollinators are important for
1Department of Forest Sciences, University of British Columbia, Vancouver, BC, Canada,
V6T 1Z4. 2Department of Zoology, University of Wisconsin-Madison, Madison, WI 53706.
3Department of Natural Sciences, Mars Hill University, Mars Hill, NC 28754. *Corresponding
author - michellegooch@gmail.com.
Manuscript Editor: Jason Cryan
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
318
crop production throughout the world (Losey and Vaughan 2006), and native bees
are as (if not more) important for crop pollination as managed or feral Apis mellifera
L. (Honeybee) (Rader et al. 2013; Winfree et al. 2007a, b). Many studies
are also beginning to examine the ways in which the presence of forests in close
proximity to agricultural fields can improve pollination of crops by native bees and
other insects (Mandelik et al. 2012, Ricketts 2004, Watson et al. 2011, Winfree et
al. 2007c). However, there exists little understanding of the importance and diversity
of pollinators within natural habitats, particularly temperate forests. Pollinator
diversity is essential to natural ecosystems because many plants are specifically
adapted for certain pollinator characteristics. Without diverse pollinator availability,
such species will be unpollinated and suffer population decline (Pimentel
1997). Southern Appalachian forests are considered to be centers of plant diversity
and endangered species (Dobson et al. 1997), yet the region’s pollinator community
on which these plants depend is largely understudied (but see Campbell et al. 2007).
Because native pollinators are ecologically important, it is essential to understand
how past and present management practices affect pollinator communities.
Clear-cut timber harvesting was historically common throughout the southern
Appalachian Mountains, resulting in a patchwork of forest stands that represent different
successional stages. Logging can negatively affect the forest understory plant
community in a variety of ways, including by reducing species richness, cover, and
abundance of native herbs (Bellemare et al. 2002, Brown and Boutin 2009, Vellend
2005, Wyatt and Silman 2010), reducing population growth rates for certain species
(Jackson et al. 2013), and increasing cover of invasive species (Kuhman et al. 2011).
Changes in pollination frequency or timing may be one mechanism responsible for
altered plant performance in logged sites. In a recent meta-analysis, Winfree et al.
(2009) found a negative effect of human disturbance on the abundance and species
richness of wild bees, but only in cases of extreme habitat loss. Some moderate
disturbances may actually help to maintain pollinator numbers and diversity due
to increased light reaching the ground, which stimulates growth of early successional
flowering plants that provide nectar and pollen resources (Romey et al. 2007,
Taki et al. 2013, Vallet et al. 2010). Increased light is also associated with higher
temperatures and irradiance that can directly influence pollinator activity and nesting
(Kilkenny and Galloway 2008, Potts and Wilmer 1997). Studies have found
positive associations between the abundance and species richness of pollinators
and disturbances that reduce the density of overstory trees and increase cover of
herbaceous plants (Campbell et al. 2007, Romey et al. 2007, Winfree et al. 2007b).
However, certain pollinator groups, including some rare endemics (Winfree et al.
2007b) and wood-nesting bees (Jha 2009), show positive associations with forested
habitat. Following logging, canopy openness, herb cover, and tree species richness
show nonlinear relationships with stand age, with all three highest in very young
stands, lowest in mid-aged stands, and increasing again in very old stands (Elliott
and Swank 1994, Jules et al. 2008). Therefore, logging might result in optimal habitat
for some pollinator groups at first, but decline in habitat suitability over time as
succession proceeds.
Southeastern Naturalist
319
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
Logging roads, embedded in otherwise continuous forest, are clear, long-lasting
remnants of timber operations. Roads create long, narrow edges that may extend
well beyond the surface of the road itself and may persist for more than 40 years
after abandonment (Vora 1988). There has been little research on the effects of logging
roads on forest biodiversity (but see Forman 2000, Haskell 2000, Semlitsch
et al. 2007) despite their prevalence in the landscape. Logging roads negatively affected
salamander populations up to 35 m into the forest on both sides of the roads
in the southern Appalachians (Semlitsch et al. 2007), and soil arthropods declined
in proximity to forest roads (Haskell 2000). However, insect pollinators may be
attracted to habitat near roads due to increased light incident on the forest floor
(Trombulak and Frissell 2000) and the additional floral resources that grow on or
adjacent to them (Vallet et al. 2010).
The southern Appalachians are characterized by steep topographic gradients,
with temperature declining at a consistent rate with increasing elevation (Bolstad
et al. 1998). Due to the importance of temperature and moisture on insect physiology
and behavior (Hodkinson 2005), elevation likely represents another important
driver of pollinator composition and abundance in this region. Insects are less
active and may be less abundant at high elevations due to lower ambient temperatures,
resulting in fewer pollinator visits (Arroyo et al. 1982). Associations between
insect abundance and elevation can be direct, with thermoregulatory capacity playing
an important role, or indirect and mediated through insects’ interactions with
other organisms including plants, competitors, parasitoids, predators, and pathogens
(Hodkinson 2005). Flies (order Diptera) are often relatively more abundant in
cooler, wetter, high-elevation sites than bees (Hymenoptera: Apoidea), which tend
to dominate at warmer, dryer, lower sites (Devoto et al. 2005, Kearns 1992).
Our objectives were to determine how pollinator abundance and community
composition at the taxonomic family level varied with stand age, distance to logging
roads, and elevation in southern Appalachian forests. We hypothesized that
pollinator community-composition would differ between old stands that were
logged >90 years ago and young stands (20–40 years post-harvest) because some
pollinator groups are better-adapted to disturbance than others (Carré et al. 2009).
We also predicted that pollinator groups that favor open forest-canopies associated
with disturbances, such as ground-nesting bees (Campbell et al. 2007) and
certain fly pollinators and parasitoids such as Syrphidae and Bombyliidae (Kwak
et al. 1996, Proctor et al. 1996), would be more abundant close to logging roads,
whereas forest-dependent groups (e.g., flies in the super-family Muscoidea and
some wood-nesting bees such as Megachilidae; Steffan-Dewenter 2002) would
increase in abundance farther from roads. Lastly, we predicted that the pollinator
community would change along an elevation gradient, with flies being relatively
more abundant than bees at higher elevations.
Methods
Study area
The study was conducted in the French Broad River Basin in western North
Carolina (Fig. 1). The region lies within the southern Blue Ridge physiographic
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
320
province, which is characterized by steep terrain and mixed-mesophytic deciduous
forest (Braun 1950). This area receives 125 cm of precipitation per year on average,
with mean winter temperatures of 4 °C and mean summer temperatures of 23 °C
(Southeast Regional Climate Center 2008). Elevation ranges from 350 m to 1900
m. Forests are dominated by northern hardwoods at higher elevations, mixed hardwoods
on less-fertile lower-elevation sites, and mixed mesophytic forests on lower
slopes and in coves (SAMAB 1996). The biological diversity of herbaceous communities
is particularly rich, especially in mesic cove forests (Glenn-Lewin 1977).
The topographic variability that characterizes the region creates strong gradients of
temperature and moisture (Bolstad et al. 1998).
This region of the southern Appalachians has a history of extensive timber harvesting
in the 1800s and early 1900s, followed by land abandonment and natural
reforestation that began in the mid-1900s and has continued to the present (Eller
1982). Although much of the land is forested today, timber harvesting continues
throughout the region, resulting in a patchwork of forest stands in different successional
stages.
Field methods
We selected 15 previously logged forest stands situated along logging roads
and along an elevation gradient. Stand age was estimated from US Forest Service
(USFS) Continuous Inventory of Stand Condition data followed by visual inspection
in the field. We categorized stands last harvested prior to 1920 as old (n = 8),
and those last harvested between 1970 and 1990 as young (n = 7). The average
width of logging roads was 4.5 m (± 1.5 m). Because logging operations were ongoing
in portions of the study area, all roads were closed to public vehicle traffic
Figure 1. Map of the 15 study sites in the French Broad River watershed (shaded area, top
right panel), western North Carolina (showing topography using the Hillshade feature in
ArcMap), and the placement of pan traps (bowl color: white = open, blue = darkest shading,
and yellow = lightest shading) along transects at 0, 2, 10, 50, and 100-m parallel to logging
roads at each site (bottom right panel).
Southeastern Naturalist
321
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
but were periodically mowed and maintained (e.g., cleared of fallen debris) by the
USFS. None of the roads we designated as sampling sites were actively utilized by
logging trucks during the 2-year study period.
We conducted insect trapping in summer (late May–July 2010) and spring (late
March–April 2011). At each site, we established 75-m transects parallel to the
logging road at 0 (edge of the road), 2, 10, 50, and 100 m from the road (Fig. 1).
Two sites lacked sufficient area for 5 transects; therefore one site only contained
four transects at 0–50 m and one had 3 at 0–10 m. We placed fifteen 355-ml disposable
plastic bowls (pan traps) on the forest floor 5 m apart along each transect.
Bowls were painted yellow, blue, and white because these colors represent common
floral colors and have been shown to attract large numbers of pollinators
(Campbell and Hanula 2007). We filled the bowls three-quarters full with water,
added unscented dishwashing detergent to reduce surface tension, and left them
out continuously for 4–15 days during each sampling season (average of 9 days
in both seasons). All bowls at a given site were left out for the same number of
days, but some sites contained bowls for fewer days than others due to issues with
accessibility. We checked all bowls every 2–3 days throughout the continuous
4–15-day study period in each season. During each check, we pooled the insects
from all bowls along a transect into a single sample and immediately transferred
the specimens to 70% isopropyl alcohol for storage, and we refilled the bowls
with soapy water. We sorted the insects and identified those known to be potential
pollinators—members of the orders Hymenoptera, Diptera, Coleoptera, and Lepidoptera—
to taxonomic family. Several other insect orders were caught in the pan
traps, but we did not consider them to be flower visitors; thus, we excluded them
from this study. We placed voucher specimens at the University of Wisconsin Insect
Research Collection, Madison, WI.
Along each transect within a site, we measured vegetation variables hypothesized
to vary with stand age, distance to road, and elevation. These included
canopy openness (measured with a handheld spherical densiometer), percent cover
of forbs and shrubs, and presence of blooming flowers. We made all vegetation
measurements within 1-m2 quadrats at 5 evenly spaced points along each transect,
and averaged the data at the transect level for both spring and summer.
Statistical analysis
To explore how vegetation variables differed with stand age, distance to road,
and elevation, we conducted one-way ANCOVA (for categorical variables) and
linear mixed-effects models (for continuous variables), with site included as a random
effect. We transformed continuous variables as necessary (either square-root
or fourth-root transformations) to achieve normality and homogeneity of variance.
Due to small sample size (n = 15 sites), we considered associations among variables
to be significant if P < 0.1. We conducted separate analyses for spring and summer.
We analyzed insect abundance data, measured as numbers of individuals per
family, using a revised approach to a multi-level modeling (MLM) framework
(Gelman and Hill 2007) as described by Jackson et al. (2012). Multilevel models
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
322
make it possible to simultaneously estimate the responses of multiple species (or
families, in this case) to environmental variables, thereby also giving a summary of
the environmental determinants of community composition (see Jackson et al. 2012
for a formal description of the model).
In this case, the MLM can be interpreted as a set of Poisson regressions in
which differences in slopes and intercepts among families are random variables.
Model fixed-effects give the average response of individual insect families to environmental
variables, thus providing information on how average abundance of
all flower visitors is affected by a particular variable. The variance of the random
effects gives the variability among families in their response to the environmental
variable (i.e., changes in community composition). The larger the variation
among families in response to an environmental variable, the greater the changes
in family-level community composition along the corresponding environmental
gradient. Additionally, the MLM simultaneously gives information at the family
level in the form of regression coefficients of the effects of environmental variables
on each family individually.
We chose a set of insect families known to contain pollinators as response
variables in the MLM. Through a comprehensive study in a temperate deciduous
forest near the study region, Motten (1986) quantified all floral visitors to the spring
wildflower community. The pollinators listed in Motten’s study comprised nine
families and three orders (Hymenoptera, Diptera, and Lepidoptera; Table 1), and
the majority were bees in the families Andrenidae, Apidae, Halictidae, and Megachilidae,
and flies in the family Bombyliidae (which also parasitize andrenid bees).
These were also common pollinating families at our study sites (M. Jackson, pers.
observ.). Therefore, we included in our analysis the pollinating families reported by
Motten (1986) that were found in >10% of our transects. The one exception to this
was that we combined Anthomyiidae and Muscidae into the super-family Muscoidea.
Due to its high diversity and the potential for incorrect identification for some
families within Muscoidea, we analyzed this super-family as one entity.
We developed one MLM for each season (spring 2011 and summer 2010) and
included elevation, stand age, and distance to road as predictor variables. We
transformed elevation by subtracting the mean and dividing that value by the
standard deviation. Stand age and distance to road were both treated as categorical
variables. We chose to treat distance to road categorically because of observed
non-linear responses of insects to distance to road that we felt were ecologically
meaningful and best captured categorically. We compared old vs. young stands,
and the 0 m transect vs. all other transect distances from the road at each site. All
predictor variables were included as fixed effects as well as random effects in the
models. We modeled insect abundance aggregated across transects as count data
using a log-normal Poisson distribution, which can be implemented by including
an individual-level random-effect term corresponding to each observation (i.e.,
transect–family combination) in the model (Agresti 2002, Elston et al. 2001). The
combined number of pan-trap hours per transect was included as an offset variable
(Bates et al. 2011) in the model to account for sampling effort. To identify important
Southeastern Naturalist
323
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
environmental effects on community composition, we tested whether variances of
predictor-variable random-effects differed from zero using likelihood-ratio tests.
Because variables were aggregated at the transect-level, we included site as a separate
random effect in the models. The MLMs were constructed using lmer (Bates et
al. 2011) in R version 2.14.1 (R Development Core Team 2011).
Lastly, to test for an interaction between stand age and distance to road, we
conducted two-way ANCOVAs with the total abundance of all pollinating insect
families as response variables and site as a random effect. Because native bees are
the dominant pollinators of the spring wildflower community (Motten 1986), we
also tested this interaction for all bees as a separate group. Abundance data were
corrected for sampling effort by dividing by trap-hours and multiplying by 100, and
were fourth-root transformed to achieve normality.
Results
Light and forest understory conditions
Sites represented a range of elevations (704–1555 m) and environmental conditions
(see Supplemental Table 1, available online at https://www.eaglehill.us/
SENAonline/suppl-files/s13-2-S2057-Jackson-s1, and, for BioOne subscribers, at
http://dx.doi.org/10.1656/S2057.s1). In old stands, the overstory was dominated by
Quercus rubra L. (Red Oak), Acer saccharum Marshall (Sugar Maple), and Carya
spp. (hickory), whereas young stands were dominated by disturbance-adapted species
such as Liriodendron tulipifera L. (Tulip-poplar), Prunus pensylvanica L. (Pin
Cherry), and Robinia pseudoacacia L. (Black Locust).
Canopy openness was higher in spring than in summer (16.5% and 7.8%, respectively;
t = -10.8, P < 0.0001) and did not vary with distance to road or stand age. In
summer, canopy openness was significantly lower at 2 m and 10 m from the road
relative to transects adjacent to the road (0 m: F = 2.36, P = 0.06). Forb and shrub
cover were lower in spring than in summer (forb: 13.3% and 37.0%, respectively;
t = 10.1, P < 0.0001; shrub: 8.8% and 14.7%, respectively; t = 3.6, P = 0.0006). In
spring, forb cover was positively correlated with elevation (R2 = 0.40, P = 0.01) and
was higher in old stands than young stands (20% and 5%, respectively; F = 3.69,
P = 0.08). Flowers were more common in spring than summer (54.2% and 28.6%,
respectively; χ2 = 10.1, P = 0.001), and their presence increased with elevation (χ2 =
8.07, P = 0.005) and forb cover (χ2 = 9.15, P = 0.002) in spring. In summer, flowers
were more common at 0 m from the road than at farther distances (χ2 = 10.45, P=
0.03), with no effect of distance to road in spring. The presence of flowers did not
vary with stand age in spring or summer.
Pollinator community
In spring, we captured a total of 13,755 insects from 67 families within 4 orders
known to contain flower visitors, and in summer, 26,739 insects from 108
families (Table 1, Supplemental Table 2, available online at http://www.eaglehill.
us/SENAonline/suppl-files/s13-2-S2057-Jackson-s1 and, for BioOne subscribers,
at http://dx.doi.org/10.1656/S2057.s1). In spring, the potential-pollinator
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
324
community captured in our samples was dominated by Hymenoptera (66.8% of
the overall captured community), particularly bees in the families Megachilidae
and Andrenidae, followed by order Diptera (29.8% of the captured community;
Table 1). In summer, Diptera was the most abundant order (60.7% of the captured
community), dominated by muscoid flies. Hymenoptera represented 25.4% of
Table 1. Abundance (insects/trap-hours × 100) and frequency (% of transects present; n = 72) of the 10
most abundant insect families captured within each of 4 orders known to contain pollinators, including
all families used in the multilevel model (denoted by A) and those found to be important pollinators by
Motten (1986; denoted by B). Insect families included in the multilevel model are shown even when
not in the 10 most abundant families within an order.
Spring 2011 Summer 2010
Abundance Frequency Abundance Frequency
Hymenoptera 320.60 100.0 Hymenoptera 297.30 100.0
MegachilidaeA, B 146.66 98.6 Diapriidae 109.89 98.7
AndrenidaeA, B 54.47 100.0 Ichneumonidae 81.98 100.0
HalictidaeA, B 32.46 88.9 HalictidaeA, B 12.39 75.3
Ichneumonidae 27.61 93.1 Pteromalidae 10.72 66.2
ApidaeA, B 18.31 90.3 Tenthredinidae 5.88 59.7
Pomphilidae 7.84 68.1 Pomphilidae 5.80 50.6
Tenthredinidae 6.97 54.2 Cynipidae 4.10 50.6
Cynipidae 2.46 43.1 AndrenidaeA, B 3.38 48.1
Diapriidae 1.56 38.9 Platygastridae 1.93 45.5
ColletidaeA 1.27 30.6 ApidaeA, B 1.70 39.0
MegachilidaeB 0.35 9.1
Diptera 138.20 100.0 Diptera 569.20 100.0
TachinidaeA, B 56.92 100.0 MuscoideaA, B 160.01 100.0
Ceratopoganidae 31.26 73.6 Dolichopodidae 149.08 98.7
Phoridae 15.34 91.7 Phoridae 56.24 100.0
Sciaridae 14.67 80.6 Chironomidae 34.13 51.9
SyrphidaeA, B 4.47 69.4 Drosophilidae 27.47 94.8
Mycetophilidae 4.38 58.3 Rhagionidae 24.30 92.2
Tipulidae 2.79 34.7 TachinidaeA, B 21.27 94.8
BombyliidaeA, B 2.68 36.1 Sciaridae 13.49 88.3
Chironomidae 1.84 25.0 Callihporidae 10.96 62.3
Ottidae 0.85 1.4 Tipulidae 9.75 68.8
SyrphidaeA, B 6.27 58.4
BombyliidaeA, B 0.02 1.3
Coleoptera 12.60 95.8 Coleoptera 118.40 100.0
Carabidae 4.79 66.7 Staphylinidae 50.84 93.5
Chrysomelidae 3.51 61.1 Nitidulidae 26.15 70.1
Staphylinidae 1.25 23.6 Carabidae 11.43 85.7
Trogidae 0.69 23.6 Chrysomelidae 8.31 77.9
Curcurlionidae 0.50 19.4 Lampyridae 7.35 48.1
Leiodidae 0.38 13.9 Scarabidae 2.43 53.2
Meloidae 0.36 11.1 Curcurlionidae 2.06 54.5
Oedemeridae 0.36 11.1 Cerambycidae 1.90 45.5
Cantharidae 0.16 6.9 Elateridae 1.62 39.0
Elateridae 0.16 5.6 Throscidae 1.43 14.3
Southeastern Naturalist
325
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
the captured community in summer, and was dominated by parasitic wasps in the
families Diapriidae and Ichneumonidae.
Spring
Neither elevation nor stand age were associated with average abundance
of flower-visiting families in spring (MLM fixed effects; Table 2). Increasing
distance to road was negatively related with average abundance of individuals;
Table 2. Multilevel model fixed and random-effect coefficients for southern Appalachian pollinator
families in spring 2011 and summer 2010. Old = an indicator of site stand age (old stands were logged
>90 years ago, and young stands were logged 20–40 years ago). RdDist = distance from the logging
road in meters (n = 3–5 transects/site); coefficients are in comparison to the 0-m transect.
Predictor Fixed estimate SE P Random estimate SD P
Spring
Intercept -6.51 0.64 <0.0001 0.00 0.00
Elevation 0.01 0.19 0.96 0.11 0.33 <0.0001
Old -0.24 0.34 0.48 0.24 0.49 <0.0001
RdDist 0.06
2 m -0.54 0.13 <0.0001 0.02 0.14
10 m -0.45 0.13 0.0004 0.00 0.01
50 m -0.37 0.13 0.005 0.00 0.06
100 m -0.33 0.16 0.04 0.07 0.27
Summer
Intercept -6.14 0.49 <0.0001 0.00 0.00
Elevation -0.14 0.17 0.43 0.09 0.31 <0.0001
Old -0.44 0.26 0.09 0.10 0.32 0.003
RdDist <0.0001
2 m -0.97 0.26 0.0002 0.31 0.56
10 m -0.94 0.24 0.0001 0.25 0.50
50 m -0.64 0.26 0.01 0.30 0.55
100 m -0.56 0.25 0.03 0.28 0.52
Table 1, continued.
Spring 2011 Summer 2010
Abundance Frequency Abundance Frequency
Lepidoptera 26.60 54.2 Lepidoptera 5.70 64.9
PieridaeB 5.96 16.7 Hesperiidae 1.82 33.8
Geometridae 3.92 13.9 Geometridae 1.16 29.9
Tortricidae 3.17 11.1 Nymphalidae 0.37 9.1
Noctuidae 1.85 6.9 Noctuidae 0.37 10.4
Lycaenidae 1.60 6.9 Papilionidae 0.18 3.9
Hesperiidae 1.39 5.6 Tortricidae 0.14 6.5
Gelechiidae 1.31 4.2 Adelidae 0.06 2.6
Nymphalidae 1.26 5.6 Gelechioidea 0.05 2.6
Papilionidae 1.04 4.2 Mimallonidae 0.03 1.3
Elachistidae 0.94 2.8 Lycaenidae 0.02 1.3
PieridaeA, B 0.02 1.3
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
326
significantly more flower-visiting insects were found adjacent to the road (0 m)
than at any other transect.
Community composition of flower visitors, as represented by among-family
variation in abundance, was significantly influenced by elevation, stand age, and
distance to road (MLM random effects; Table 2). Using a threshold coefficient of
|0.3| to define variable importance when random effects were extracted by families,
2 bee families were more abundant at high elevations in spring (Andrenidae and
Apidae; order Hymenoptera), and Bombyliidae (order Diptera) was more abundant
at low elevations (Table 3). Abundance of 4 families across 3 orders was higher in
young forests in spring (Halictidae, Megachilidae, Bombyliidae, and Pieridae), but
Andrenidae was more abundant in old forests. In spring, all families were more
abundant at 0 m from roads than at 2 m, but only 4 were more abundant at 0 m than
at 100 m, suggesting a positive but diminishing effect of roads as distance to road
increased.
We found no significant interaction between stand age and distance to road on
total pollinator abundance in a two-way ANCOVA. In spring, bees, as a subgroup,
were reduced in abundance at the 2-m transect (distance to road: F = 4.7, P = 0.003;
Fig. 2), consistent with the results from the MLM.
Summer
Average abundance of summer flower-visiting insects was greater in young forests
than in old forests, and closer to roads (Table 2).
Table 3. Multilevel model random-effect coefficients for elevation, old (an indicator of site stand age;
see Table 2), and RdDist2-100 (distance from logging road in meters; coefficients are in comparison to
the 0-m transect) extracted by pollinator family for spring 2011 and summer 2010; asterisks indicate
coefficients >|0.3|. Coefficients are the random effects plus the estimate for fixed effects in order to
account for the mean slope.
Distance from logging road
Elevation Old 2 m 10 m 50 m 100 m
Spring
Andrenidae 0.41* 0.44* -0.63* -0.44* -0.32* -0.25
Apidae 0.39* 0.08 -0.57* -0.44* -0.35* -0.50*
Colletidae 0.07 -0.11 -0.39* -0.45* -0.43* -0.24
Halictidae -0.11 -0.36* -0.63* -0.44* -0.33* -0.36*
Megachilidae -0.10 -0.82* -0.75* -0.44* -0.27 -0.11
Bombyliidae -0.46* -0.53* -0.44* -0.45* -0.41* -0.19
Syrphidae -0.27 -0.16 -0.48* -0.45* -0.39* -0.34*
Tachinidae 0.06 0.10 -0.67* -0.44* -0.31* -0.70*
Pieridae 0.10 -0.69* -0.36* -0.45* -0.44* -0.30*
Summer
Andrenidae 0.20 -0.36* -1.38* -1.31* -1.04* -0.95*
Apidae -0.28 -0.42* -1.60* -1.51* -1.26* -1.16*
Halictidae -0.01 -0.81* -0.83* -0.81* -0.50* -0.43*
Muscoidea 0.12 -0.04 0.01 -0.05 0.32* 0.36*
Syrphidae -0.57* -0.34* -1.19* -1.14* -0.86* -0.78*
Tachinidae -0.22 -0.63* -0.61* -0.62* -0.29 -0.23
Southeastern Naturalist
327
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
All predictor variables significantly influenced community composition of flower-
visiting insects in summer. All families were more abundant in young forests and
close to roads in summer except for Muscoidea (order Diptera), which showed little
association with stand age but was more abundant far from roads. The negative effect
of increasing distance to road for all families except Muscoidea and Tachinidae
was strong out to 100 m. Elevation was less important as a driver of individual family
abundances in summer than in spring, although Syrphidae (order Diptera) was
more abundant at low elevations.
In summer, bees, as a subgroup, were strongly affected by distance to road in
young forests (they were more abundant at 0 m), with no effect of distance to road
in old forests (interaction term: F = 3.2, P = 0.02; Fig. 2).
Discussion
Our results show that pollinator community composition and factors determining
insect abundance in this region varied substantially between the spring and
summer. Habitat associated with logging roads and recently cut stands appeared to
enhance pollinator abundances for open and edge-loving taxa. Increased pollinator
abundance in younger stands and in close proximity to roads was more pronounced
in summer (when canopy was closed) than in spring, when the canopy was open.
The benefits of roads to pollinator abundance appeared to diminish as forests matured
and old-growth characteristics (i.e., canopy gaps) were restored. Elevation
was less important than logging history and was more variable in its effects on the
pollinator community.
Logging legacies
Conditions in recently logged stands. Differences in the pollinator community
across stand age, including greater average abundance of pollinators in young
forests in summer, were likely driven by altered environmental conditions that
augmented nesting and food resources in recently logged stands. Forb cover was
Figure 2. Mean number (± 1 SE) of bees corrected for sampling effort (trap-hours-1*100)
caught at varying distances to road (n= 72 transects in spring, 60 transects in summer) in old
versus young stands in a) spring (late March–April, 2011) and b) summer (late May–July,
2010).
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
328
lower in young stands in spring, and in a concurrent study in the same region,
Jackson et al. (2013) found a shallower litter layer in young stands. All of the bee
families in our samples typically nest in the ground and are dependent on bare soil
(Potts et al. 2010), although some species of Apidae, Halictidae, and Megachilidae
nest in wood. With shallower litter and reduced forb cover, the nesting conditions
for bees were likely enhanced in young stands. However, we found no effect of
stand age on overall insect or bee abundance in spring, indicating that habitat
conditions in young stands, other than litter depth and forb cover, were also suitable
for pollinators. The greater importance of stand age for predicting pollinator
abundance in summer may have been driven by altered species composition of
summer-blooming plants, potentially including higher species richness and greater
prevalence of competitive, wind-dispersed plant species at the expense of slowergrowing,
mature-forest indicator species (Brown and Boutin 2009, Romey et al.
2007). Indeed, based on an extensive review, Roulston and Goodell (2011) suggest
that floral resource availability is the primary factor influencing wild bee abundance
and that little evidence supports nest-site limitation. Although we did not measure
plant species composition in our stands, we speculate that the presence of competitive
flowering species following recent logging may provide enhanced nectar
resources for pollinators, as described by Romey et al. (2007).
Effects of logging roads. In spring, abundances for many pollinator families
resembled a U-shaped distribution relative to distance to road; abundances were
lower at middle distances (2-10 m), but increased to numbers similar to the 0-m
transect by 100 m into the forest. Neither canopy openness nor the presence of
flowers were significantly lower 2-m from roads in spring, but high shrub-cover
resulted in a denser micro-environment that may have deterred floral-visiting
insects from foraging in those areas. Flowers (and the pan traps) may have been
harder for pollinators to detect at mid-distances where vegetative cover was higher.
Still, some families were reduced in abundance even at 100 m in spring; these
included Apidae, many species of which are reportedly uncommon in woodlands
(Williams 1988) and attracted to clearings (Proctor et al. 1996). Because none of
the spring vegetation or light measurements were affected by roads at 100 m, we
suggest that the preference of these families for roads may have been driven partially
by unmeasured variables such as plant-species composition or homogeneity
of canopy openness.
In summer, all families except Muscoidea and Tachinidae declined sharply
in abundance with increasing distance from roads all the way out to 100 m into
the forest. Canopy openness was significantly reduced at 2–10 m from roads, but
gradually increased to levels similar to those found adjacent to roads by 100 m.
Canopy gaps are associated with increased pollinator visits (Totland et al. 2005,
Walters and Stiles 1996), likely by enhancing conditions for nesting (Potts and
Wilmer 1997) and/or foraging (Kilkenny and Galloway 2008). Furthermore, edge
effects associated with woodland roads include increased light, temperature, and
floral availability, which create favorable conditions for some pollinators (Vallet
et al. 2010, Wright et al. 2010). Fewer woodland plants flower in mid-summer,
Southeastern Naturalist
329
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
and those that do are typically associated with clearings (Proctor et al. 1996).
Indeed, more flowers were blooming adjacent to roads in summer than spring,
suggesting that nectar resources primarily determined pollinator abundance. Muscoid
flies were the only group to decline in abundance closer to logging roads;
however, Muscoidea is an extremely diverse superfamily that includes some
species that do not visit flowers at all, making it difficult to draw definitive conclusions
regarding Muscoidea as a whole.
Habitat associated with logging roads only affected summer bee abundance
(families Andrenidae, Apidae, and Halictidae combined) in young forests, suggesting
that effects of roads may diminish as forests mature. Previous work in the
region has shown that old forests contain more canopy gaps than younger forests
(Jackson et al. 2013). Therefore, the relative importance of the road as a canopy gap
in summer may be higher in young forests because bees in old forests have access
to a greater abundance of small canopy gaps whether or not roads are present. In
spring, the canopy was homogenously open across all stands; thus pollinator abundance
was driven less by effects of logging on light availability and resulting floral
resources and more by other factors (e.g., shrub cover). The phenology of bees may
also play a large role in determining these relationships between bee abundance,
season, stand age, and distance to road. Many species of andrenid bees emerge in
early spring in deciduous forests (Motten 1986), but there is great diversity in the
family Andrenidae. In addition to phenologic differences, some species may respond
differently to the presence of roads, making it impossible to draw definitive
conclusions about the mechanisms underlying this interaction between stand age
and distance to road across seasons.
Effects of elevation
The effects of elevation were most pronounced in spring (March–April), which
is the bloom time for many wildflower species in the region. Two important groups
of bee pollinators, Andrenidae and Apidae, were more abundant at high elevations.
Bombyliidae (the bee flies) declined in abundance with elevation in spring.
These results contradicted our hypothesis that bees would be more abundant at
low elevations, and flies would be more common at high elevations. Bombus spp.
(bumblebees, family Apidae) are better-adapted for cooler temperatures than most
bees because their large bodies and dense pelage provide increased thermoregulatory
capacity (Bishop and Armbruster 2002); Pyke (1982) also found reduced
bumblebee abundance at low elevations along transects in Colorado. However,
more than half of our Apidae specimens from spring collections were in the genus
Nomada and not Bombus (indeed, pan traps often capture fewer bumblebees than
expected based on their observed abundance; Roulston et al. 2007), so Bombus was
likely not the driver of greater Apidae abundance at high elevations. Additionally,
forb cover and flower presence increased with elevation, perhaps due to a moisture
gradient. Therefore, increased abundance of Andrenidae and Apidae at high elevations
may have been driven by a combination of thermoregulatory capacity and
available nectar resources. In contrast, diversity and abundance of bombyliid flies
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
330
(as well as bee diversity, though not in this study) have been shown to decrease with
elevation (Arroyo et al. 1982, Kearns 1992) and peak in arid and semi-arid environments
(Hull 1973), which are more typical of low-elevation sites in this region. In
general, bee diversity tends to be higher in arid and semi-arid environments than in
mesic environments.
Management implications
The positive response of many pollinators to logging legacies indicates that
pollination is not likely a mechanism driving reduced performance of native forest
understory herbs in recently logged sites (Jackson et al. 2013), at least in this
system. Indeed, management practices that temporarily open the forest canopy can
likely be used to positively influence pollinator abundance. For instance, Romey
et al. (2007) found that bee abundance and diversity was highest in logged stands
where the most trees had been removed. Logging roads create long-lasting gaps in
the forest canopy, particularly when they are maintained after logging operations
have ceased. Roads fragment forest habitat and can negatively affect some taxa
(Haskell 2000, Ortega and Capen 1999, Semlitsch et al. 2007). However, our results
show that logging roads may actually serve to enhance understory habitat for
certain pollinators, at least temporarily. We caution against assuming that logging
leads to a long-term increase in bee abundance and diversity because we found that
the benefits associated with logging for native bees were diminished in later successional
(i.e., older) forests. Still, logging roads may provide supplemental light
and/or floral resources in the short term, until canopy-gap dynamics associated with
older forests are reestablished.
Despite the apparent short-term benefits for some pollinator groups, negative
effects of logging on understory-herb diversity can last for centuries (Duffy and
Meier 1992, Wyatt and Silman 2010), and mismatched responses to anthropogenic
disturbances could ultimately lead to a decoupling of plant–pollinator interactions
(Potts et al. 2010). Alternatively, pollinators may exhibit complementary habitat
use, which allows certain species and communities to persist in highly dynamic
systems by seasonally moving to different patches within a landscape. For example,
Mandelik et al. (2012) found that when summed over time, bee communities in two
different habitats (agricultural and old fields) contained similar species, but habitat
use by specific species and by the community as a whole changed through the
season. Bees, in particular, are highly mobile organisms adapted to using patchy,
seasonally changing resources (Potts et al. 2010), although mobility is highly dependent
on body size (Gathmann and Tscharntke 2002). In our system, pollinators
may move between forest patches and, to a more limited extent, across elevations,
to access preferable floral and nesting resources as they change across seasons.
We cannot make strong inferences about seasonal pollinator movement because
our study only spanned a single summer and spring; further multi-year studies are
needed to verify our speculations about seasonality.
We must also acknowledge the limitations of pan traps as our sole mode of insect
sampling. Pan traps are often used because of their efficiency, lack of observer
Southeastern Naturalist
331
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
bias, and cost effectiveness (Baum and Wallen 2011, Westphal et al. 2008). We
chose this method in order to efficiently sample several remote locations over a
wide geographic area (Cane et al. 2000). However, some researchers have shown
that pan traps are biased toward certain pollinators (e.g., small bees such as halictids),
and fail to capture others (e.g., larger-bodied bees such as bumblebees and
carpenter bees) (Cane et al. 2000, Roulston et al. 2007, Wilson et al. 2008). Our
comparisons were all among sites sampled with the same methodology, so any bias
was consistent across treatments. However, our findings might have been different
if we had employed another sampling technique (e.g., sweep netting). Placing pan
traps on the ground also may have biased sampling toward certain groups because
some insects might only encounter traps if they are placed at the same height as
their preferred forage flowers (Cane et al. 2000). Nearly all of the observed flowers
blooming during the two sampling periods were on low-lying forbs rather than
shrubs or trees, so we believe this bias to be minimal. A final caveat is that pollinator
presence at flowers is likely dependent on flower density and time since peak
flowering (Lazaro and Totland 2010, Linsley 1978). We captured some information
regarding flower density in our data on the presence of flowers along each transect.
Furthermore, we sampled during the peak wildflower bloom-time in the southern
Appalachians (early April) and in the summer when a different plant community
was blooming. Therefore, given the capture biases inherent in pan trapping, we
likely captured the greatest possible diversity of pollinators in our sites.
Despite the importance of pollinators for maintaining native forest-plant diversity,
understanding of their habitat needs and responses to disturbances remains
limited. Logging continues today in the southern Appalachians, and roads represent
long-lasting legacies of these operations. Climate warming, one of the greatest
threats to biodiversity worldwide, will likely interact with land use to affect species
distributions in ways that are difficult to predict (Thuiller et al. 2006). Apid and andrenid
bees, comprising the dominant spring pollinator families in Motten’s (1986)
and our studies, were more abundant at high elevations and thus may be at elevated
risk of decline under climate warming because they will have less available habitat
for future movement upslope. The potential for interacting effects of climate and
land use on species distributions requires that researchers consider multiple factors
simultaneously in order to better understand how pollinators and their host plants
will respond to global change. Future research should quantify the effects of land
use and environmental gradients on plant–pollinator interactions in this region,
including visitation rates and pollinator effectiveness for focal-plant species.
Acknowledgments
This study was funded by the Long-term Ecological Research Program of the National
Science Foundation (DEB-0823293 Coweeta LTER) and a Bunde Fund Research Grant
from the Department of Zoology, University of Wisconsin-Madison, WI. Kyle Johnson,
Steven Krauth, Phil Pellitteri, and Dan Young provided valuable assistance in identification
of insect specimens, and Anthony Ives helped with the statistical analyses. We appreciate
comments from Anthony Ives, Don Waller, Tom Givnish, and Jack Williams on earlier
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
332
versions of the manuscript, and for helpful suggestions from two anonymous reviewers. We
thank Jaimie Little, Aki Masunaga, Adam Milch, Bryan Moore, Chris Platt, Klara Rossouw,
and Abigail Kohl for their field and lab assistance, and the US Forest Service for a permit
to conduct research in the Pisgah National Forest.
Literature Cited
Agresti, A. 2002. Categorical Data Analysis, 2nd Edition. Wiley Press, Hoboken, NJ.
Aizen, M., P. Feinsinger, G. Bradshaw, and P. Marquet. 2003. Bees not to be? Responses of
insect-pollinator faunas and flower pollination to habitat fragmentation. Pp. 111–130, In
G. Bradshaw and P. Marquet (Eds.). How Landscapes Change: Human Disturbance and
Ecosystem Fragmentation in the Americas. Springer-Verlag, Berlin, Germany.
Arroyo, M., R. Primack, and J. Armesto. 1982. Community studies in pollination ecology
in the high temperate Andes of central Chile. I. Pollination mechanisms and altitudinal
variation. American Journal of Botany 69:82–97.
Bates, D.M., M. Maechler, and B.M. Bolker. 2011. lme4: Linear mixed-effects models using
S4 classes, Version 0.999375-39. R Foundation for Statistical Computing, Vienna,
Austria. Available online at http://cran.r-project.org/web/packages/lme4/. Accessed 14
January 2013.
Baum, K.A., and K.E. Wallen. 2011. Potential bias in pan trapping as a function of floral
abundance. Journal of the Kansas Entomological Society 84:155–1 59.
Bellemare, J., G. Motzkin, and D.R. Foster. 2002. Legacies of the agricultural past in the
forested present: An assessment of historical land-use effects on rich mesic forests.
Journal of Biogeography 29:1401–1420.
Biesmeijer, J.C. 2006. Parallel declines in pollinators and insect-pollinated plants in Britain
and the Netherlands. Science 313:351–354.
Bishop, J.A., and W.S. Armbruster. 2002. Thermoregulatory abilities of Alaskan bees: Effects
of size, phylogeny, and ecology. Functional Ecology 13:711–724.
Bolstad, P.V., L. Swift, F. Collins, and J. Régnière. 1998. Measured and predicted air temperatures
at basin to regional scales in the southern Appalachian mountains. Agricultural
and Forest Meteorology 91:161–176.
Bond, W. 1994. Do mutualisms matter? Assessing the impact of pollinator and disperser
disruption on plant extinction. Philosophical Transactions of the Royal Society of London.
Series B: Biological Sciences 344:83–90.
Braun, E. 1950. Deciduous Forests of Eastern North America. McGraw-Hill, New York, NY.
Brown, C.D., and C. Boutin. 2009. Linking past land-use, recent disturbance, and dispersal
mechanism to forest composition. Biological Conservation 142:1647–1656.
Campbell, J.W., and J.L. Hanula. 2007. Efficiency of Malaise traps and colored pan-traps
for collecting flower-visiting insects from three forested ecosystems. Journal of Insect
Conservation 11:399–408.
Campbell, J., J. Hanula, and T. Waldrop. 2007. Effects of prescribed fire and fire surrogates
on floral-visiting insects of the Blue Ridge province in North Carolina. Biological Conservation
134:393–404.
Cane, J.H., and V.J. Tepedino. 2001. Causes and extent of declines among native North
American invertebrate pollinators: Detection, evidence, and consequences. Conservation
Ecology 5(1): 1. Available online at http://www.consecol.org/vol5/iss1/art1/. Accessed
12 September 2012.
Cane, J., R. Minckley, and L. Kervin. 2000. Sampling bees (Hymenoptera: Apiformes) for
pollinator-community studies: Pitfalls of pan-trapping. Journal of the Kansas Entomological
Society 73:225–231.
Southeastern Naturalist
333
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
Carré, G., P. Roche, R. Chifflet, and N. Morison. 2009. Landscape context and habitat type
as drivers of bee diversity in European annual crops. Agriculture, Ecosystems, and Environment
133:40–47.
Devoto, M., D. Medan, and N.H. Montaldo. 2005. Patterns of interaction between plants
and pollinators along an environmental gradient. Oikos 109(3):461–472.
Dobson, A., J. Rodriguez, W. Roberts, and D. Wilcove. 1997. Geographic distribution of
endangered species in the United States. Science 275:550–553.
Duffy, D.C., and A.J. Meier. 1992. Do Appalachian herbaceous understories ever recover
from clearcutting? Conservation Biology 6:196–201.
Eller, R. 1982. Miners, Millhands, and Mountaineers: Industrialization of the Appalachian
South, 1880–1930. University of Tennessee Press, Knoxville, TN.
Elliott, K., and W. Swank. 1994. Changes in tree-species diversity after successive clearcuts
in the southern Appalachians. Plant Ecology 115:11–18.
Elston, D.A., R. Moss, T. Boulinier, and C. Arrowsmith. 2001. Analysis of aggregation, a
worked example: Numbers of ticks on Red Grouse chicks. Parasito logy 122:563–569.
Forman, R. 2000. Estimate of the area affected ecologically by the road system in the United
States. Conservation Biology 14:31–35.
Gathmann, A., and T. Tscharntke. 2002. Foraging ranges of solitary bees. The Journal of
Animal Ecology 71:757–764.
Gelman, A., and J. Hill. 2007. Data Analysis Using Regression and Multilevel/Hierarchical
Models. Cambridge University Press, New York, NY.
Glenn-Lewin, D. 1977. Species diversity in North American temperate forests. Plant Ecology
33:153–162.
Haskell, D. 2000. Effects of forest roads on macroinvertebrate soil fauna of the southern
Appalachian mountains. Conservation Biology 14:57–63.
Hodkinson, I.D. 2005. Terrestrial insects along elevation gradients: Species and community
responses to altitude. Biological Reviews 80:489–513.
Hull, F.M. 1973. Bee flies of the world. The genera of the family Bombyliidae. Bulletin of
the United States National Museum No. 286. Smithsonian Institution Press, Washington,
DC.
Jackson, M.M., M.G. Turner, S.M. Pearson, and A.R. Ives. 2012. Seeing the forest and
the trees: Multilevel models reveal both species and community patterns. Ecosphere
3:art79. Available online at http://esapubs.org/archive/ecos/C003/C_03.htm. Accessed
12 October 2012.
Jackson, M.M., S.M. Pearson, S.M., M.G. Turner. 2013. Performance and population dynamics
of a native understory herb differ between young and old forest stands in the
Southern Appalachians. Forest Ecology and Management 304:444–454.
Jha, S. 2009. Movement in the Matrix: Pollination and dispersal processes in a tropical
coffee and forest landscape mosaic. Ph.D. Dissertation. University of Michigan, Ann
Arbor, MI.
Jules, M.J., J.O. Sawyer, and E.S. Jules. 2008. Assessing the relationships between stand
development and understory vegetation using a 420-year chronosequence. Forest Ecology
and Management 255:2384–2393.
Kearns, C.A. 1992. Anthophilous fly distribution across an elevation gradient. American
Midland Naturalist 127:172–182.
Kearns, C.A., and D.W. Inouye. 1997. Pollinators, flowering plants, and conservation biology.
BioScience 47:297–307.
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
334
Kearns, C.A., D.W. Inouye, and N.M. Waser. 1998. Endangered mutualisms: The conservation
of plant-pollinator interactions. Annual Review of Ecology and Systematics
29:83–112.
Kilkenny, F.F., and L.F. Galloway. 2007. Reproductive success in varying light environments:
Direct and indirect effects of light on plants and pollinators. Oecologia,
155(2):247–255.
Kremen, C., and T. Ricketts. 2000. Global perspectives on pollination disruptions. Conservation
Biology 14:1226–1228.
Kuhman, T.R., S.M. Pearson, and M.G. Turner. 2011. Agricultural land-use history increases
non-native plant invasion in a southern Appalachian forest a century after abandonment.
Canadian Journal of Forest Research 41:920–929.
Kwak, M.M., O. Velteron, and J.M. Els Boerrigter. 1996. Insect diversity and the pollination
of rare plant species. Pp. 115–124, In A. Matheson, S.L. Buchmann, C. O'Toole, P.
Westrich, and I.H. William (Eds.). The Conservation of Bees. Linnean Society Symposium
Series, vol. 18. Academic Press. 254 pp.
Lazaro, A., and Ø Totland. 2010. Local floral composition and the behaviour of pollinators:
Attraction to and foraging within experimental patches. Ecological Entomology
35:652–661.
Linsley, E.G. 1978. Temporal patterns of flower visitation by solitary bees, with particular
reference to the Southwestern United States. Journal of the Kansas Entomological Society
51:531–546.
Losey, J.E., and M. Vaughan. 2006. The economic value of ecological services provided by
insects. BioScience 56:311.
Mandelik, Y., R. Winfree, T. Neeson, and C. Kremen. 2012. Complementary habitat use by
wild bees in agro-natural landscapes. Ecological Applications 22:1535–1546.
Memmott, J., P.G. Craze, N.M. Waser, and M.V. Price. 2007. Global warming and the disruption
of plant-pollinator interactions. Ecology Letters 10:710–717.
Motten, A. 1986. Pollination ecology of the spring wildflower-community of a temperate
deciduous forest. Ecological Monographs 56:21–42.
Ollerton, J., R. Winfree, and S. Tarrant. 2011. How many flowering plants are pollinated by
animals? Oikos, 120:321–326.
Ortega, Y.K., and D.E. Capen. 1999. Effects of forest roads on habitat quality for Ovenbirds
in a forested landscape. The Auk 116:937–946.
Pimentel, D., C. Wilson, C. McCullum, R. Huang, P. Dwen, J. Flack, et al. 1997. Economic
and Environmental Benefits of Biodiversity. BioScience 47:747–757.
Potts, S., and P. Willmer. 1997. Abiotic and biotic factors influencing nest-site selection
by Halictus rubicundus, a ground-nesting halictine bee. Ecological Entomology
22:319–328.
Potts, S.G., J.C. Biesmeijer, C. Kremen, and P. Neumann. 2010. Global pollinator declines:
Trends, impacts, and drivers. Trends in Ecology and Evolution 25:345–353.
Proctor, M., P. Yeo, and A. Lack. 1996. The Natural History of Pollination. Timber Press,
Portland, OR.
Pyke, G.H. 1982. Local geographic distributions of bumblebees near Crested Butte, Colorado:
Competition and community structure. Ecology 63:555–573.
R Development Core Team. 2011. R: A language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria.
Rader, R., J. Reilly, I. Bartomeus, and R. Winfree. 2013. Native bees buffer the negative impact
of climate warming on Honeybee pollination of watermelon crops. Global Change
Biology 19:3103–3110.
Southeastern Naturalist
335
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
Ricketts, T. 2004. Tropical forest fragments enhance pollinator activity in nearby coffee
crops. Conservation Biology 18:1262–1271.
Romey, W.L., J.S. Ascher, D.A. Powell, and M. Yanek. 2007. Impacts of logging on midsummer
diversity of native bees (Apoidea) in a northern hardwood forest. Journal of the
Kansas Entomological Society 80:327–338.
Roulston, T.H., and K. Goodell. 2011. The role of resources and risks in regulating wild bee
populations. Annual Review of Entomology 56:293–312.
Roulston, T.H., S.A. Smith, and A.L. Brewster. 2007. A comparison of pan-trap and intensive
net-sampling techniques for documenting a bee (Hymenoptera: Apiformes) fauna.
Journal of the Kansas Entomological Society 80:179–181.
Semlitsch, R.D., T.J. Ryan, K. Hamed, M. Chatfield, B. Drehman, N. Pekarek, M. Spath,
and A. Watland. 2007. Salamander abundance along road edges and within abandoned
logging roads in Appalachian forests. Conservation Biology 21:159–167.
Southeast Regional Climate Center. 2008. Southeast Regional Climate Center, Station
310301. Asheville, NC.
Southern Appalachian Man and the Biosphere (SAMAB). 1996. The southern Appalachian
assessment summary report. US Department of Agriculture Forest Service, Southern
Region, Atlanta, GA.
Steffan-Dewenter, I. 2002. Landscape context affects trap-nesting bees, wasps, and their
natural enemies. Ecological Entomology 27:631–637.
Taki, H., I. Okochi, K. Okabe, T. Inoue, H. Goto, T. Matsumura, and S. Makino. 2013.
Succession influences wild bees in a temperate forest landscape: The value of early
successional stages in naturally regenerated and planted forest s. PLoS ONE 8:e56678.
Thuiller, W., D.M. Richardson, M. Rouget, and S. Proches. 2006. Interactions between environment,
species traits, and human uses describe patterns of plant invasions. Ecology
87:1755–1769.
Totland, Ø., P. Nyeko, A.L. Bjerknes, and S.J. Hegland. 2005. Does forest-gap size affect
population size, plant size, reproductive success, and pollinator visitation in Lantana
camara, a tropical invasive shrub? Forest Ecology and Management 215: 329–338.
Trombulak, S., and C. Frissell. 2000. Review of ecological effects of roads on terrestrial
and aquatic communities. Conservation Biology 14:18–30.
Vallet, J., V. Beaujouan, J. Pithon, and F. Rozé. 2010. The effects of urban or rural landscape
context and distance from the edge on native woodland-plant communities. Biodiversity
and Conservation 19:3375–3392.
Vellend, M. 2005. Land-use history and plant performance in populations of Trillium grandiflorum.
Biological Conservation 124:217–224.
Vora, R. 1988. Potential soil compaction forty years after logging in northeastern California.
Western North American Naturalist 48:117–120.
Walters, B.B., and E.W. Stiles. 1996. Effect of canopy gaps and flower patch-size on pollinator
visitation of Impatiens capensis. Bulletin of the Torrey Botanical Club 123:184–188.
Warren, M.S., J.K. Hill, J.A. Thomas, J. Asher, R. Fox, B. Huntley, D.B. Roy, M.G. Telfer,
S. Jeffcoate, P. Harding, G. Jeffcoate, S.G. Willis, J.N. Greatorex-Davies, D. Moss, and
C.D. Thomas. 2001. Rapid responses of British butterflies to opposing forces of climate
and habitat change. Nature 414:65–69.
Watson, J.C., A.T. Wolf, and J.S. Ascher. 2011. Forested landscapes promote richness and
abundance of native bees (Hymenoptera: Apoidea: Anthophila) in Wisconsin apple orchards.
Environmental Entomology 40:621–632.
Southeastern Naturalist
M.M. Jackson, M.G. Turner, and S.M. Pearson
2014 Vol. 13, No. 2
336
Westphal, C., R. Bommarco, G. Carre, E. Lamborn, N. Morison, T. Petanidou, S.G. Potts,
S.P.M. Roberts, H. Szentgyrgyi, T. Tscheulin, B.E. Vaissiere, M. Woyciechowski, J.C.
Biesmeijer, W.E. Kunin, J. Settele, and I. Steffan-Dewenter. 2008. Measuring bee diversity
in different European habitats and biogeographical regions. Ecological Monographs
78:653–671.
Williams, P.H. 1988. Habitat use by bumble bees (Bombus spp.). Ecological Entomology
13:223–237.
Wilson, J.S., T. Griswold, and O.J. Messinger. 2008. Sampling bee communities (Hymenoptera:
Apiformes) in a desert landscape: Are pan traps sufficient? Journal of the Kansas
Entomological Society 81:288–300.
Winfree, R., N. Williams, J. Dushoff, and C. Kremen. 2007a. Native bees provide insurance
against ongoing honey bee losses. Ecology Letters 10:1105–1113.
Winfree, R., T. Griswold, and C. Kremen. 2007b. Effect of human disturbance on bee communities
in a forested ecosystem. Conservation Biology 21:213–223.
Winfree, R., N.M. Williams, H. Gaines, J.S. Ascher, and C. Kremen. 2007c. Wild bee pollinators
provide the majority of crop visitation across land-use gradients in New Jersey
and Pennsylvania, USA. Journal of Applied Ecology 45:793–802.
Winfree, R., R. Aguilar, D.P. Vázquez, G. LeBuhn, and M.A. Aizen. 2009. A meta-analysis
of bees’ responses to anthropogenic disturbance. Ecology 90:2068–2076.
Wright, T.E., S. Kasel, M. Tausz, and L.T. Bennett. 2010. Edge microclimate of temperate
woodlands as affected by adjoining land use. Agricultural and Forest Meteorology
150:1138–1146.
Wyatt, J.L., and M.R. Silman. 2010. Centuries-old logging legacy on spatial and temporal
patterns in understory herb communities. Forest Ecology and Management 260:116–124.