Bee Abundance and Nutritional Status in 
Relation to Grassland Management 
Practices in an Agricultural Landscape
Authors: Griffin W. Smith, Diane M. Debinski, Nicole 
A. Scavo, Corey J. Lange, John T. Delaney, 
Raymond A. Moranz, James R. Miller, David M. 
Engle, and Amy L. Toth
This is a pre-copyedited, author-produced PDF of an article accepted for publication in 
Environmental Entomology following peer review. The version of record "Smith, Griffin W., Diane 
M. Debinski, Nicole A. Scavo, Corey J. Lange, John T. Delaney, Raymond A. Moranz, James R. 
Miller, David M. Engle, and Amy L. Toth. 2016. Bee Abundance and Nutritional Status in Relation 
to Grassland Management Practices in an Agricultural Landscape. Environmental Entomology, 
45(2), 338–347. Available at: http://dx.doi.org/10.1093/ee/nvw005" is available online at: https://
doi.org/10.1093/ee/nvw005.
Smith, Griffin W., Diane M. Debinski, Nicole A. Scavo, Corey J. Lange, John T. Delaney, 
Raymond A. Moranz, James R. Miller, David M. Engle, and Amy L. Toth. 2016. Bee Abundance 
and Nutritional Status in Relation to Grassland Management Practices in an Agricultural 
Landscape. Environmental Entomology, 45(2), 338–347. Available at: http://dx.doi.org/10.1093/
ee/nvw005.
Made available through Montana State University’s ScholarWorks 
scholarworks.montana.edu 
Bee Abundance and Nutritional Status in Relation to
Grassland Management Practices in an Agricultural
Landscape
Griffin W. Smith,1 Diane M. Debinski,1 Nicole A. Scavo,1 Corey J. Lange,1 John T.
Delaney,1 Raymond A. Moranz,1 James R. Miller,2 David M. Engle,3 and Amy L. Toth1,4
1Department of Ecology, Evolution, and Organismal Biology, Iowa State University, Ames, IA 50011 (griffinwsmith@gmail.com;
debinski@iastate.edu; nicole.a.scavo@gmail.com; coreyjlange@gmail.com; johntdelaney@gmail.com; raymoranz@yahoo.com;
amytoth@iastate.edu), 2Department of Natural Resources and Environmental Sciences, University of Illinois, Urbana, IL 61801
(jrmillr@illinois.edu), 3Department of Natural Resource Ecology and Management, Oklahoma State University, Stillwater, OK
74078 (david.engle@okstate.edu), and 4Corresponding authors, e-mail: amytoth@iastate.edu
Received 26 August 2015; Accepted 19 January 2016
Abstract
Grasslands provide important resources for pollinators in agricultural landscapes. Managing grasslands with
fire and grazing has the potential to benefit plant and pollinator communities, though there is uncertainty about
the ideal approach. We examined the relationships among burning and grazing regimes, plant communities,
and Bombus species and Apis mellifera L. abundance and nutritional indicators at the Grand River Grasslands
in southern Iowa and northern Missouri. Treatment regimes included burn-only, grazed-and-burned, and patch-
burn graze (pastures subdivided into three temporally distinct fire patches with free access by cattle). The
premise of the experimental design was that patch-burn grazing would increase habitat heterogeneity, thereby
providing more diverse and abundant floral resources for pollinators. We predicted that both bee abundance
and individual bee nutritional indicators (bee size and lipid content) would be positively correlated with floral re-
source abundance. There were no significant differences among treatments with respect to bee abundance.
However, some of the specific characteristics of the plant community showed significant relationships with bee
response variables. Pastures with greater abundance of floral resources had greater bee abundance but lower
bee nutritional indicators. Bee nutritional variables were positively correlated with vegetation height, but, in
some cases, negatively correlated with stocking rate. These results suggest grassland site characteristics such
as floral resource abundance and stocking rate are of potential importance to bee pollinators and suggest
avenues for further research to untangle the complex interactions between grassland management, plant
responses, and bee health.
Key words: grassland management, patch-burn grazing, honey bee, bumble bee, pollinator nutrition
Intensive agriculture across large geographic areas has greatly
altered the landscape and forage plants available for both native and
managed pollinators (Kremen et al. 2002, National Research
Council 2006, Ricketts et al. 2008, Winfree et al. 2009). Remnants
of native grasslands within an agricultural landscape have the poten-
tial to provide excellent food resources and nest sites for pollinators.
However, grasslands may be managed in several different ways,
leading to different plant communities, which can affect the abun-
dance and health of local insect communities (Hines and Hendrix
2005; Moeller 2005; Davis et al. 2007, 2008; Slagle and Hendrix
2009; Hendrix et. al 2010). It is thus important to understand the
ecological factors within grasslands that are most relevant to the
abundance and health of bee pollinators.
Grassland vegetation characteristics such as nectar availability,
pollen resources, and vegetation height have the potential to affect
bees in a variety of ways. Vegetation height affects the distribution
of nesting spots and refuges. Availability and diversity of floral re-
sources can affect bee nutritional status. Foraging in landscapes with
limited nectar and pollen sources has the potential to lead to nutri-
tional deficits, which in turn may influence bees’ susceptibility to
fungal, bacterial, and viral diseases (Myack and Naug 2009, Di
Pasquale et al. 2013). Bumble bee colonies closer to patches of rich
floral resources produce more workers, healthier larvae, and larger
individuals on average (Sutcliffe and Plowright 1990, Pereboom
2000, Pelletier and McNeil 2003, Westphal et al. 2009).
Conversely, colonies next to abundant resources may also
experience higher rates of parasitism, negating some of the positive
nutritive benefits (Carvell et al. 2008). Thus, a better understanding
of the connections between grassland management, vegetation re-
sponses, and bee health is needed.
One opportunity for enhancing pollinator resources in the
United States lies in the management of native (i.e., unplowed)
Midwestern grasslands that are currently used for cattle grazing.
Here we examine how fire and grazing management practices within
grasslands affect the abundance and nutritional state of common
bee species: native bumble bees (genus Bombus) and the nonnative
honey bee, Apis mellifera L. We examined bees foraging in pastures
managed using three different treatments: burn-only, graze-and-
burn, and patch-burn graze (pastures subdivided into three tempor-
ally distinct fire units (patches) with free access by cattle). This work
is part of an experiment to assess the effects of grassland manage-
ment practices on vegetation and wildlife in the Grand River
Grasslands, working agricultural grasslands of southern Iowa and
northern Missouri (Debinski et al. 2011, Miller et al. 2012). This re-
gion is characterized by rolling hills that are low in corn suitability,
and thus the predominant land use is cattle grazing. Most of the
grazing is done at a high stocking rate, leaving very little in the way
of vegetation structure in the pastures. However, there has been a
concerted effort to protect and restore prairies in the region, so the
landscape is a mosaic of grazed pastures, ungrazed grasslands, and
rowcrops. Given the high potential for restoration, this region is an
ideal location for testing how plant community characteristics and
management techniques can affect the health and abundance of both
wild and managed bees.
Returning two key and interconnected processes (fire and graz-
ing) to native grasslands has the potential to enhance pollinator
habitat, although there is still uncertainty and some controversy re-
garding the ideal approach to using grazing and burning for inverte-
brate conservation (Swengel 1998; Panzer and Schwartz 2000;
Cook and Holt 2006; Vogel et al. 2007, 2010; Engle et al. 2008;
Moranz 2010; Swengel et al. 2011). Burning stimulates flower pro-
duction, which increases resources for pollinators (Platt et al. 1988,
Rudolph et al. 2006, Brewer et al. 2009). Cattle grazing can increase
plant species richness in grasslands, especially through preserving
the diversity of forb species (Harrison et al. 2003, Hayes and Holl
2003). However, homogeneous application of grazing or burning
can lead to less diverse wildlife (Holecheck et al. 2003, Kirchner
et al. 2011) and insect communities (Reed 1997, Branson et al.
2006). Alternatively, a combination of fire and grazing has been
proposed to maintain and restore essential habitat heterogeneity
(Fuhlendorf and Engle 2001). The habitat heterogeneity provided by
patch-burn grazing has the potential to provide more diverse and
abundant floral resources for pollinators (Fuhlendorf and Engle
2001, Helzer and Steuter 2005). However, relatively few researchers
have assessed the impacts of patch-burn-grazing on insect commun-
ities (Engle et al. 2008; Debinski et al. 2011; Doxon et al. 2011;
Moranz et al. 2012, 2013, 2014), and none have considered the nu-
tritional health of pollinators.
We measured plant community characteristics and sampled
A. mellifera and several Bombus species from grazed, burned-and-
grazed, and patch-burn-grazed grasslands. We then examined bee
nutritional indicators to determine if bee nutrition was related to
treatments, plant community characteristics, or both. Nutritional in-
dicators included head width and wing length (both indicators of
larval nutrition), lipid content (an indicator of adult nutrition), and
abdomen mass (an indicator of both larval and adult nutrition;
Winston 1987). We also estimated whether species abundance var-
ied among pastures as a function of treatment or plant community
characteristics. Our hypotheses were as follows: 1) patch-
burn-grazed treatments will lead to higher floral plant diversity and
average floral resource abundance (FRA), 2) pastures providing
greater floral resources will have more bees with higher nutritional
indicators, and 3) pastures with heaviest grazing will support
smaller numbers of bees with lower nutritional indicators.
Materials and Methods
Study Sites
We delineated 12 pastures in the Grand River Grasslands (see map
in Supp. Fig. 1 [online only]) in 2006. All pastures are located in a
160 square kilometer radius in Ringgold County of southern Iowa
and Harrison County of northern Missouri. These pastures have
served as study sites in previous studies, assessing the effectiveness
of patch-burn grazing in improving habitat for grassland insects and
birds (Moranz et al. 2010, Debinski et al. 2011, Pillsbury et al.
2011, Hovick et al. 2012). The pastures range in size from 15.6 to
37.4 ha and are a mix of public and private grasslands. For more in-
formation on the plant community of the study, see McGranahan
et al. (2013). Pastures were allocated to one of the three treatments:
1) patch-burn-graze—annual burning of spatially distinct patches
and free access by cattle, 2) graze-and-burn—burning entire tracts
with free access by cattle, and 3) burn-only—burning of entire pas-
tures with no grazing (typical management for protected lands in the
region). The burning schedule was on a three-year cycle in all cases,
and the number of cattle (stocking rate) was low. From 2007
through 2009, pastures in grazing treatments were stocked with cat-
tle at a rate of 3.4 animal unit months (AUM, the amount of for-
age consumed per animal unit per month, where an animal unit is
defined as a 1,000-pound cow and her suckling calf) per ha between
May 1 and October 1. The stocking rate was reduced to about 1.5
AUM per ha from 2010–2012 to maximize the treatment goal of
creating heterogeneity among patches within the pastures. Each pas-
ture was divided into three patches of approximately equal area. In
patch-burn graze pastures, natural topographic features such as
waterways, drainages, and ridgetops were used as patch boundaries
to the extent possible. Each year, a different patch within each
patch-burn graze pasture was burned in early spring. Burn-only and
graze-and-burn pastures were burned in their entirety every three
years. All pastures or patches were burned during mid-March of the
burn years within approximately a two-week window of time.
During 2006–2009, all 12 pastures were used for the study (four
pastures per treatment). During 2011–2012, only six pastures were
used in the study (two pastures per treatment). See the Supp. Table 1
(online only) for a complete list of pastures and vegetation and bee
sampling methods used each year. Most sites were separated by >1
mile; however, some sites were closer together (Supp. Fig. 1 [online
only]) and, thus, may have partially shared bee communities. While
honey bee and bumble bees may forage greater than 1 mile from
their colonies (Beekman and Ratnieks 2001), many foraging flights
occur closer to nests (Steffan-Dewenter and Kuhn 2003). Thus, we
expect fairly distinct bee communities foraged on each of our sites.
Quantification of Vegetation Variables
We quantified three plant community features: 1) vegetation height
using a Robel pole (quantified as in Debinski et al. 2011), 2) average
FRA, and 3) number of flowering plants. Robel data were collected
annually during the peak of the growing season in early July; vegeta-
tion height was measured to provide an indicator or the level of
growth of a stand of grassland after burning or intensive grazing
(Moranz et al. 2012). We expected higher vegetation height to be
associated with more floral resources, as there is little vegetation left
beyond a “grazing lawn” 1 cm in height in intensively grazed pas-
tures, providing no floral resources (Debinski, personal observa-
tion). In addition, FRA and diversity data were collected twice
during summer (the first round between June 7–July 3, second round
between July 5–August 5) starting in 2008. These dates were not ne-
cessarily the same dates on which bees were collected; instead, the
two measurements were taken to capture overall trends in floral re-
sources, not necessarily the floral resources available to bees at the
time of capture. Two 100-m transects were established within each
patch (three per pasture). Within each of the five 1- by 20-m seg-
ments within a transect, the number of flowering ramets were
counted for each flowering plant species. We summed the total num-
ber of ramets by species within each transect and then averaged
across transects for each pasture. We calculated floral plant diversity
using Shannon’s Diversity Index (Magurran 2004). While these
measurements of FRA and diversity do not provide specific informa-
tion about the actual availability of nectar and pollen to bees, they
do provide a proxy for the amount and diversity of potential forage
available to pollinators. Note that data on several additional vegeta-
tion variables were collected as part of another related project
(Debinski et al. 2011): cover of warm season grasses, cool season
grasses, tall fescue (Schedonorus arundinaceus (Schreb.), nonlegu-
minous forbs, leguminous forms, and woody species. We analyzed
these data for relationships with bee abundance and nutrition but
did not find meaningful correlations and, therefore, decided not to
use them in further analyses.
Bee Collections for Nutritional Indicators
From 2006 to 2009, we collected A. mellifera and Bombus, repre-
senting several species (described in results), from 12 pastures solely
to be used for analysis of nutritional indicators (described below).
We collected twice each summer—once from June to early July and
once from mid-July to early August—with an observer (the same ob-
server across all sites within a year) using a nontargeted sweep net-
ting (i.e., not specifically aiming to capture bees) along a 50-m
transect. The number of net sweeps (40) rather than the number of
minutes was standardized in this effort (see Debinski et al. 2011 for
complete 2006–2009 methodology).
In 2011, we surveyed bees from a subset (6) of pastures, provid-
ing two replicates of each management treatment (see Supp. Table 1
[online only]). These bees were again only used for measurement of
nutritional indicators, thus, times and sweeps were not standardized
among the pastures. We collected once in June, July, and August for
1.5-h periods. For these collections, a collector walked around the
pastures, stopping at open floral blooms and netting in a targeted
way in order to collect A. mellifera and Bombus species. Bombus
were identified to species after collection (in the laboratory) using
publicly available online keys (Ballew and Pickering 2002).
Bee Collections for Estimating Abundance
In 2012, we collected Apis and Bombus species in late June and
early August of 2012 from the same six grassland pastures as the
2011 collections. Bee abundance in the August collection was abnor-
mally low due to a prolonged drought and was therefore excluded
from further analysis. A collector (G.W.S.) captured bees by walking
three established 100-m transects in each pasture at a pace of 20 m/
30 s. Bees spotted 10 m ahead of or alongside the collector were
captured via sweep netting. Each of the three surveys lasted for
10 min with the clock stopped for capture and handling of bees in
order to identify them to species. Afterwards, bees were released.
For all years, surveys were limited to times when the temperature
was between 21C and 35C, wind was <16 km/h, clouds did not
obscure the sun, and between 0900 and 1800 hours. Although bee
abundance and activity can vary greatly throughout the day, we
were unable to finely control for the time of day of bee sampling due
to logistical challenges of traveling to all of our sites.
Nutritional Indicators
We identified collected bees according to species, sex, and queen or
worker caste. Only female workers were used for further analysis.
To estimate overall bee size, we measured head width and wing
length by photographing each specimen and measuring these fea-
tures digitally using the Leica Application Suite (version 2.0.0). To
estimate wet mass, we removed abdomens from each bee and
weighed them on a microbalance. Bombus species have substantial
variation in worker body size (up to 10-fold in body mass,
Couvillon et al. 2010), whereas body size variation in A. mellifera
workers is much lower. Nonetheless, A. mellifera worker body size
can be significantly affected by nutritional stress during larval or
pupal development (Brodschneider and Crailsheim 2010), thus, sug-
gesting that such body size measurements had the potential to be in-
formative for A. mellifera as well. We conducted lipid analysis via a
colorimetric assay and estimated total lipid content of individual bee
abdomens based on a standard curve of pure cholesterol (see Toth
and Robinson 2005). Each species was analyzed separately. Wing
length and head width data were available for all five years (2006–
2009 and 2011), but lipid and abdomen mass data were only avail-
able for bees collected in 2011. Lipid stores are known to vary sig-
nificantly across the lifetime of honey bee workers, with foragers
having lower lipid stores than nonforagers (Toth and Robinson
2005). Nonetheless, these studies suggest there is some additional
loss of lipid after the onset of foraging and variation in lipid stores
between foragers (Toth and Robinson 2005); thus, examining lipid
stores of foragers can still provide information about nutritional sta-
tus of foragers.
Statistical Analyses
A complete list of vegetation and bee measures used each year is
available in Supp. Table 1 (online only).
Vegetation Data
We tested for correlations among continuous vegetation and pasture
variables with JMP 10’s multivariate feature using data for all years
in which bees were collected (JMP Version 10.0.0). We then used
ANOVA to test the relationship between treatment and all of the
continuous vegetation and pasture variables. Note that we included
data for all pastures where bee surveys were conducted, even if the
bee abundance was zero.
Bee Data
We analyzed bee data with JMP software using three statistical
approaches. First, we ran all variables—as model effects—against
bee abundance and nutritional variables, using standard least
squares models with an effect leverage emphasis and a restricted
maximum likelihood (REML) method with pasture included as a
random effect. For some of the nutrition relationships, the sample
size was prohibitively small, preventing JMP from running the least
squares model, and in these cases, we ran an ANOVA with no ran-
dom effects (for treatment, the only discrete predicting variable) and
a bivariate regression analysis (for the four continuous predicting
variables: AUM, vegetation height, FRA, and floral plant diversity)
to test for relationships between vegetation variables and bee abun-
dance or nutritional variables. We compared treatments using post
hoc t-tests with a Bonferroni correction. We examined the responses
of the total bee community, A. mellifera, all Bombus species com-
bined (“total Bombus”), B. griseocollis and B. impatiens (the two
most abundant Bombus species), and all other Bombus species
(those that were not B. griseocollis or B. impatiens). Sample sizes
were limiting, especially for analyses at the species level; however,
we decided to analyze species separately due to concerns about
lumping together bee species of different sizes and life histories and
because we thought it would be informative to determine whether
multiple species showed similar responses to treatment variables.
Note that, because the study involved multiple years of data and site
variables such as AUM and vegetation variables changed from year
to year, we always used data from each pasture and for each year,
and did not use any average values in our statistical analysis.
Results
Correlations Among Pasture Variables
To better understand the relationships between vegetation and
stocking rate, we examined pairwise correlations across all continu-
ous vegetation variables (floral plant diversity, vegetation height,
FRA), in addition to examining correlations of each vegetation vari-
able with AUM, across the six pastures. The correlations among the
variables ranged widely (from r¼0.90 to 0.54), and were all
highly significant (Supp. Fig. 2 [online only]). Higher FRA was cor-
related with lower floral plant diversity, lower vegetation height,
and higher cattle stocking (AUM). Higher AUM was also correlated
with lower vegetation height and lower floral plant diversity
(Supp. Fig. 2 [online only]).
Relationships Between Treatment and Pasture
Variables
We tested for differences among the three treatments for the four
pasture variables (AUM, floral plant diversity, vegetation height,
and FRA) for all years in which bees were surveyed. The three vege-
tation variables (Fig. 1) showed significant differences among the
treatments (P<0.05). Vegetation height (ANOVA P<0.0001) and
floral plant diversity (ANOVA P¼0.0411) both showed the same
trend, having the highest means under burn-only conditions, the se-
cond highest under graze-and-burn, and the lowest under patch-
burn graze. For FRA, the pattern was the opposite (ANOVA
P¼0.0208, Fig. 1).
Bee Abundance
Bees were found on five of the six pastures sampled in 2012 (Supp.
Table 1 [online only]). Abundance among the pastures was unevenly
distributed, with collections counts ranging from 0 to 16 with an
average of 4.5 bees collected per pasture. Of the 27 bees collected in
2012, 14 were Apis and 13 were Bombus species; B. impatiens and
B. griseocollis were the two most abundant. The “other” Bombus
included B. pennsylvanicus, B. vagans, B. bimaculatus, and two spe-
cimens that were unidentified Bombus sp.
Only one of the six bee abundance measures (abundance of
“other” Bombus) showed a near significant difference among treat-
ments (Table 1); however, the sample size was very small (n¼6,
P¼0.0741). An average of 1.5 other Bombus species were found on
patch-burn graze pastures, 1.0 at burn-only pastures, and 0 at graze-
and-burn pastures.
Average FRA from 2008 to 2012 positively correlated with
abundance of several different bee groups (data shown are from
2012 only). Pastures with greater average FRA had higher overall
bee counts and more Apis and B. griseocollis (P<0.05), and there
were near significant positive trends for the total number of Bombus
and “other” Bombus species with FRA (P<0.10). Note that these
results were strongly affected by one pasture, Ringgold South, which
had a higher average FRA and more bees than the other pastures.
With Ringgold South excluded, none of the relationships were sig-
nificant. AUM, vegetation height, and floral plant diversity showed
no significant or near significant relationships with any of the bee
abundance measurements (P>0.10).
Nutritional Indicators
Bees used in the analysis of nutritional indicators were collected
from 2006–2009 and 2011. Note that A. mellifera yielded sufficient
samples for analysis in all four years, but B. impatiens and B. griseo-
collis only yielded sufficient samples in 2011. A. mellifera collected
before 2011 only have wing length and head width data, whereas all
bees collected during 2011 also have lipid content and abdomen
mass data. A total of 137 bees were found in eight different pastures;
some sampled pastures yielded zero specimens. We sampled 83
A. mellifera, 32 B. griseocollis, and 22 B. impatiens.
B. griseocollis had lower abdomen mass in patch-burn grazed
pastures (P¼0.0242, F¼6.6307, Table 2, Fig. 2). A. mellifera
showed a similar trend, with lowest lipid stores in patch-burn grazed
pastures, but this was significant only when using an ANOVA
model (P¼0.0017, F¼7.2351, Supp. Table 2 [online only], Fig. 2).
Bee taxa did not differ for any of the other nutritional indicators
(P>0.10) among treatments.
Cattle stocking rate (AUM) correlated negatively with A. melli-
fera wing length (P¼0.0282, R2¼0.0652) and head width
(P¼0.0206, R2¼0.1150, Fig. 3). There were no significant correl-
ations between AUM and nutritional variables for any other species,
but many of the other relationships also showed negative (though
nonsignificant) trends.
Vegetation height (Robel pole height) negatively correlated with
A. mellifera abdomen mass (without pasture as a random variable)
(P¼0.0082; pastures with greater vegetation heights had Apis indi-
viduals with smaller abdomen masses [Table 2]).
The average FRA in a pasture negatively correlated with several
nutritional variables for B. impatiens: wing length (P¼0.0467),
head width (P¼0.0383), and abdomen mass (P¼0.0409; Fig. 4).
FRA was not correlated with nutritional indicators of other bee spe-
cies or groups of species.
Floral plant diversity did not correlate with any of the nutritional
measures for the three species (P>0.10).
Discussion
This study provides new information about the potential impacts of
grassland management on vegetation traits relevant to bee pollin-
ators. Our data provide several new observations that suggest av-
enues for further study. First, we confirmed that management
strategies involving burning, grazing, and patch-burn grazing can af-
fect several, highly correlated vegetation variables (Debinski et al.
2011, Moranz et al. 2012), some of which have the potential to be
relevant to bee nutritional status. Second, our results verify (Morse
1980, Pleasants 1981) that FRA is important to the abundance of
honey bees and some species of bumble bees observed foraging on
flowers in those pastures, but surprisingly, high FRA in a pasture
does not necessarily mean bees will have higher nutritional state.
Third, our data suggest that bee nutritional state may depend on
other habitat considerations beyond FRA; habitat traits such as
stocking rate and vegetation height were associated with some
honey bee nutritional variables. Finally, our data show no clear rela-
tionship between diversity of flowering plants and bee abundance or
nutritional state.
Returning to our first hypothesis that patch-burn-grazing treat-
ments would lead to higher floral plant diversity and FRA, we found
only partial support. Although patch-burn-grazed pastures had
Fig. 1. Different land management treatments result in significant differences in (A) vegetation height, (B) FRA, and (C) floral plant diversity (Shannon’s diversity
index) of vegetation. P-values for the ANOVA for each variable are shown below the variable name, and different letters above the bars represent significant dif-
ferences in post hoc tests. n¼ 4 sites per treatment.
Table 1. Table of test statistics from standard least square models run with pasture as a random effect for 2012
Bee group Sample
size
Treatment Stocking rate
(AUM)
Vegetation
height
Floral resource
abundance
Floral plant
diversity
Total Bees 27 P¼ 0.5390 P¼ 0.9626 P¼ 0.8524 P5 0.02511 P¼ 0.3777
F¼ 0.7649 F¼ 0.0025 F¼ 0.0394 F¼ 12.1996 F¼ 0.9824
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
A. mellifera 14 P¼ 0.5514 P¼ 0.9564 P¼ 0.8372 P5 0.00951 P¼ 0.2793
F¼ 0.7308 F¼ 0.0034 F¼ 0.0480 F¼ 21.7756 F¼ 1.5636
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
Bombus total 13 P¼ 0.5267 P¼ 0.8273 P¼ 0.8840 P5 0.0887!1 P¼ 0.5909
F¼ 0.8000 F¼ 0.0542 F¼ 0.0242 F¼ 5.0130 F¼ 0.3405
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. impatiens 3 P¼ 0.6037 P¼ 0.3537 P¼ 0.5006 P¼ 0.6995 P¼ 0.6772
F¼ 0.60 F¼ 1.0989 F¼ 0.5469 F¼ 0.1721 F¼ 0.2009
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. griseocollis 4 P¼ 0.5630 P¼ 0.8958 P¼ 0.8565 P5 0.01471 P¼ 0.3192
F¼ 0.70 F¼ 0.0195 F¼ 0.0372 F¼ 16.9400 F¼ 1.2915
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
“Other” Bombus 6 P50.0741!# P¼ 0.6413 P¼ 0.4831 P5 0.0600!1 P¼ 0.1928
F¼ 7.00 F¼ 0.2532 F¼ 0.5962 F¼ 6.7599 F¼ 2.4465
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
Significant and near significant results are bolded and cells are shaded.þ and - signs indicate if the relationship is negative or positive for significant results,
# indicates nominal data (treatment), and ! indicates near significant P-values (greater than 0.05 but less than 0.10).
Table 2. Table of test statistics from standard least square models run with pasture as a random effect and in ANOVA runs for applied to nu-
tritional measures for 2006–2009 and 2011
Bee measure Treatment Stocking rate
(AUM)
Vegetation
height
Floral resource
abundance
Floral plant
diversity
A. mellifera wing length P¼ 0.7870 P50.0284- P¼ 0.3030 P¼ 0.6373 P¼ 0.1602
F¼ 0.2521 F¼ 6.2475 F¼ 1.0941 F¼ 0.2282 F¼ 4.5195
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
A. mellifera head width P¼ 0.5895 P50.0206- P¼ 0.1144 P¼ 0.2428 P¼ 0.3702
F¼ 0.5933 F¼ 7.7616 F¼ 2.6627 F¼ 1.5386 F¼ 1.3333
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
A. mellifera abdomen mass P¼ 0.5454 P¼ 0.2725 P50.0082^- P¼ 0.2147 P¼ 0.1342
F¼ 0.7549 F¼ 1.6825 F¼ 7.5473 F¼ 2.5371 F¼ 4.7226
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
A. mellifera lipid content P¼ 0.5626 P¼ 0.7080 P¼ 0.9702 P¼ 0.8507 P¼ 0.8883
F¼ 0.7374 F¼ 0.1695 F¼ 0.0016 F¼ 0.0422 F¼ 0.0233
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. griseocollis wing length P¼ 0.1738^ P¼ 0.7872 P¼ 0.6325 P¼ 0.1125 P¼ 0.2080
F¼ 1.8642 F¼ 0.0883 F¼ 0.3740 F¼ 2.6781 F¼ 1.9668
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. griseocollis head width P¼ 0.3679^ P¼ 0.6811 P¼ 0.4943 P¼ 0.2424 P¼ 0.4161
F¼ 1.0365 F¼ 0.7351 F¼ 0.4791 F¼ 1.4238 F¼ 1.4735
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. griseocollis abdomen mass P50.0242^ P¼ 0.7375 P¼ 0.8912 P¼ 0.1287 P¼ 0.2658
F¼ 6.6307 F¼ 0.1518 F¼ 0.0228 F¼ 5.9532 F¼ 2.3508
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. griseocollis lipid content P¼ 0.6991 P¼ 0.3946 P¼ 0.8389 P¼ 0.6193 P¼ 0.9052^
F¼ 0.7390 F¼ 1.0179 F¼ 0.0707 F¼ 0.2847 F¼ 0.0145
df¼ df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. impatiens wing length P¼ 0.6050 P¼ 0.5072 P¼ 0.9670 P50.0467- P¼ 0.2654
F¼ 0.6172 F¼ 0.5982 F¼ 0.0021 F¼ 7.4313 F¼ 1.5604
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. impatiens head width P¼ 0.5570 P¼ 0.5040 P¼ 0.9028 P50.0383- P¼ 0.2326
F¼ 0.7516 F¼ 0.6038 F¼ 0.0179 F¼ 8.6756 F¼ 1.7723
df¼ 2 df¼ 1 df¼ 1 df¼ 1 df¼ 1
B. impatiens lipid content NA P¼ 0.1586 P¼ 0.1586 P¼ 0.1586 P¼ 0.1586
F¼ 15.4502 F¼ 15.4502 F¼ 15.4502 F¼ 15.4502
df¼ 1 df¼ 1 df¼ 1 df¼ 1
AUM—animal unit months, a measure of cattle stocking rate. Significant and near significant results are bolded and shaded in grey.þ and - signs indicate if the
relationship is negative or positive for significant, continuous results. ^ notes relationships that could not be run with standard least squares models and are
ANOVA results run without pasture as a random effect. ! indicates near significant P-values (greater than 0.05 but less than 0.10); NA denotes “not available”
because of too few observations for analysis.
Fig. 2. (A) Lipid stores were significantly lower in A. mellifera (n¼55) collected in patch-burn-grazed pastures compared to other grassland management treat-
ments. (B) Abdomen weight was lowest in patch-burn graze pastures for B. griseocollis (n¼10). Note that P-values shown reflect an ANOVA model for A. melli-
fera and a least squares model for B. griseocollis.
Fig. 3. Scatter plot for a measure of cattle stocking rate (animal unit months¼AUM) plotted against bee nutritional variables. ** indicates significant correlation
(P<0.05). n¼8 pastures and 127 collected bees.
Fig. 4. Scatter plot for average FRA plotted against bee nutritional variables. ** indicates significant correlation (P<0.05). n¼8 pastures and 127 collected bees.
higher FRA, they also had lower floral plant diversity. Patch-burn-
graze pastures generally also had lower average vegetation height.
Patch-burn grazing is intended to create heterogeneous vegetation
height among patches within the pasture, so this result is not surpris-
ing (Fuhlendorf and Engle 2001, Helzer and Steuter 2005). Burning
at least a portion of the pasture annually may stimulate more flow-
ering on an annual basis than burning at a 3-yr interval. We found
that the average FRA was higher on patch-burn grazed pastures,
despite the fact that the flowering plant diversity was lower.
Interestingly, two measures of bee nutrition (A. mellifera lipid con-
tent and B. griseocollis abdomen mass) were lowest on patch-burn
grazed pastures. These responses hint at the potential for an inverse
relationship between the nutritional value of the floral resources
stimulated by patch-burn grazing and A. mellifera and B. griseocollis
nutritional measures, but additional research is needed to better
understand this pattern. It is possible that more flowers do not ne-
cessarily translate into a higher abundance of appropriate floral spe-
cies or nutritionally rich floral resources for bees.
With regard to the three treatments, there is an important caveat
with respect to legacy effects. Due to constraints of pasture owner-
ship and previous management, there were differential levels of graz-
ing applied to pastures prior to our experiment (Debinski et al.
2011, Moranz et al. 2012). These land-use legacies may have influ-
enced our ability to detect treatment effects. For that reason, the rest
of our discussion emphasizes the results of bee responses to vegeta-
tion characteristics more than to treatment effects. It is also import-
ant to note that the vegetation measurements and bee collections
were made on different dates; thus, the associations we report are
not meant to reflect immediate associations between floral resources
and bee measures. Rather, these measures are meant to serve as
broad, annual indicators of overall patterns in bee and floral re-
source presence.
Returning to our second hypothesis, that grasslands providing
greater floral resources would have more bees that exhibited positive
nutritional indicators, we again find only partial support. There was
a significant positive correlation between FRA and bee abundance
for A. mellifera and Bombus griseocollis. However, one site
(Ringgold South) had extremely high FRA compared to other sites,
and this site strongly drove the positive relationship. Considering
the impact of this single pasture, our data suggest that flowering
ramet densities above 60 are most effective at attracting bees. For
nutritional indicators, we found no relationship between FRA and
nutritional state for most bee groups. Counterintuitively, B. impa-
tiens in pastures with high FRA actually had lower nutritional par-
ameters (weight, head width, and wing length).
It is possible that high overall FRA does not necessarily mean
more or better nutrition for bees. Perhaps bees could have been at-
tracted to pastures with many flowers, but the nectar and pollen re-
sources they encountered may have proved unsuitable. Indeed,
patch-burn grazed pastures had higher FRA but lower flowering
plant diversity. Thus, these pastures could have been depauperate in
bee-preferred flower species. For example, with Bombus impatiens,
it is possible that the negative relationship between FRA and nutri-
tional state relates to a lack of a particular floral species or set of
species that are of importance for this bee species.
Second, perhaps there are complex interspecific interactions that
may result when FRA is high. Davis et al. (2008) found an inverse
relationship between bee and butterfly diversity in similar
Midwestern grasslands. Competition both within pollinator taxa
and among pollinator taxa is an area of research that warrants add-
itional study.
Third, as previously noted, other studies have found that colo-
nies in closer proximity to rich floral resources experience higher
rates of parasitism (Carvell et al. 2008). Thus, bees found at pastures
with higher quality nectar resources could also face a higher density
of parasitism, affecting nutritional indicators.
Fourth, it is possible that nutritionally challenged bees had dis-
persed from neighboring patches, in search of floral resources.
Foraging flight distances for both Apis and Bombus can occasionally
be extensive (up to 10 km, Beekman and Ratnieks 2001). Given the
size of our experimental pastures (15.6–37.4 ha), it is possible that
in some cases, bees with nests outside of our pastures visited to for-
age. Locations with especially high FRA would be likely to draw
more bees from poorer quality neighboring areas (Morse 1980).
Bees have excellent spatial memories and can return to profitable
foraging sites repeatedly over many days (Winston 1987), and honey
bees can recruit their nestmate workers to high-quality food patches
using the famous waggle dance (von Frisch 1965, Beekman and Lew
2007).
With respect to our third hypothesis, we did find evidence that
less intense grazing was associated with higher bee nutritional state
but not bee abundance. Grazing can affect both plant and insect
communities in grasslands; previous studies have often found nega-
tive associations (Rambo and Faeth 1999; Kruess and Tscharntke
2001, 2002) between grazing and the abundance of bees and other
insects. Although we found no differences in bee abundance, ours is
the first study to suggest a negative impact of grazing on bee nutri-
tional state. Further work on the nutritional quality of available flo-
ral resources is necessary to understand whether this pattern may
relate to quality or quantity of available forage.
We were somewhat surprised that the floral plant diversity did
not show a relationship with bee abundance or nutritional measures.
However, we used the Shannon diversity index, which may be too
simplistic a measure. It may be more important to look in more de-
tail at specific indicators of the true nectar and pollen resources
available to bees. Moranz et al. (2012) found that although some
Grand River Grasslands pastures had a large number of flowering
species, plant abundance was dominated by a small number of spe-
cies (Trifolium repens, Lotus corniculatus, Erigeron strigosus,
Pycnanthemum tenuifolium, and Trifolium pratense). Three out of
five of these are exotic species that may not be suitable for native
Bombus, and many of the additional species that were observed in
some pastures occurred at particularly low abundances compared to
other Midwestern prairies. Thus, presence of additional species that
may not provide good forage to bees could inflate the diversity index
without providing substantial additional resources.
Another important finding from this study is the fact that we
saw many similar patterns in bee responses for both A. mellifera and
several Bombus species. Not only do these species occupy different
ecological niches, Apis is nonnative and the Bombus species are na-
tive, but most strikingly, it is likely that the Apis sampled may have
come from nearby, managed hives, whereas Bombus are wild and
nonmanaged. There is great variation in how honey bee hives are
managed by beekeepers with respect to supplemental feeding, antibi-
otic treatments, and parasitic Varroa mite treatments, among many
other factors (VanEngelsdorp and Meisner 2010). We did not at-
tempt to control for any of this variation in our study, and such col-
ony or site-level variation may certainly have influenced our honey
bee results. Despite this, we found correlated responses of both
honey bees and bumble bees with regard to the following measures:
1) higher FRA was correlated with higher bee abundance, 2) higher
nutritional indicators were associated with lower grazing intensity,
and 3) there was not a strong relationship between bee abundance
and bee nutritional state. These data suggest that landscape and
vegetation effects can affect both wild and managed bees in similar
ways.
In summary, our data suggest that grassland management prac-
tices and some of their effects on vegetation and floral resources
have the potential to impact the abundance and nutritional state of
bees. In general, increasing floral resources will attract more bees,
and patch-burn-grazing did have higher FRA, suggesting this treat-
ment can be effective in attracting bees. However, our nutritional
data were collected from bees foraging on these pastures that may
have just been “visitors,” and thus the interpretation of the result
must be considered with caution. In general, we found positive nu-
tritional effects of lower grazing intensity, which is characteristic of
burn-only treatments. Thus, there is no single best land management
strategy to increase both bee abundance and nutritional state. High
levels of grazing are likely to be detrimental to both bee abundance
and nutrition. The results from this work provide useful insights for
future studies and similar conservation efforts throughout the agri-
cultural Midwest—a critical area for bee pollinators (Grixti et al.
2009).
Supplementary Data
Supplementary data are available at Environmental Entomology online.
Acknowledgments
We would like to thank Laura Winkler for assistance with bee collections,
Shannon Rusk for assistance at the field sites, and members of the Toth la-
boratory for comments that improved the manuscript. We also acknowledge
the support of the Iowa Agriculture and Home Economics Experiment
Station.
References Cited
Ballew, S., and J. Pickering. 2002. Bumblebees – identification guide –
Discover Life. (www.discoverlife.org/mp/20q?guide¼Bumblebees) (ac-
cessed 18 December 2014).
Beekman, M., and F. L. W. Ratnieks. 2001. Long-range foraging by the
honey-bee, Apis mellifera L. Funct. Ecol. 14: 490–496.
Beekman, M., and J. B. Lew. 2007. Foraging in honeybees—when does it pay
to dance? Behav. Ecol. 19: 255–261.
Branson, D., A. Joern, and G. Sword. 2006. Sustainable management of insect
herbivores in grassland ecosystems: New perspectives in grasshopper con-
trol. BioScience 56: 743–755.
Brewer, J. S., A. L. Cunningham, T. P. Moore, R. M. Brooks, and J. L.
Waldrup. 2009. A six-year study of fire-related flowering cues and coexist-
ence of two perennial grasses in a wet longleaf pine (Pinus palustris) sa-
vanna. Plant Ecol. 200: 141–154.
Brodschneider and Crailsheim. 2010. Nutrition and health in honey bees.
Apidologie 41: 278–294.
Carvell, C., P. Rothery, R. F. Pywell, and M. S. Heard. 2008. Effects of re-
source availability and social parasite invasion on field colonies of Bombus
terrestris. Ecol. Entomol. 33: 321–327.
Cook, W. M., and R. D. Holt. 2006. Fire frequency and mosaic burning effects
on a tallgrass prairie ground beetle assemblage. Biodivers Conserv. 15:
2301–2323.
Couvillon, M. J., J. M. Jandt, N. Duong, and A. Dornhaus. 2010. Ontogeny of
worker body size distribution in bumble bee (Bombus impatiens) colonies.
Ecol. Entomol. 35: 424–435.
Davis, J. D., D. M. Debinski, and B. J. Danielson. 2007. Local and landscape
effects on the butterfly community in fragmented Midwest U.S.A. prairie
habitats. Landsc. Ecol. 22: 1341–1354.
Davis, J. D., S. Hendrix, D. M. Debinski, and C. J. Hemsley. 2008. Butterfly,
bee and forb community composition and cross-taxon incongruence in tall-
grass prairie fragments. J. Insect Conserv. 12: 69–79.
Debinski, D. M., R. A. Moranz, J. T. Delaney, J. R. Miller, D. M. Engle, L. B.
Winkler, D. A. McGranahan, R. J. Barney, J. C. Trager, A. L. Stephenson,
et al. 2011. A cross-taxonomic comparison of insect responses to grassland
management and land-use legacies. Ecosphere 2: 131.
Di Pasquale, G., M. Salignon, Y. Le Conte, L. P. Belzunces, A. Decourtye, A.
Kretzschmar, S. Suchail, J. L. Brunet, and C. Alaux. 2013. Influence of pol-
len nutrition on honey bee health: Do pollen quality and diversity matter?.
PLoS ONE 8: e72016.
Doxon, E. D., C. A. Davis, S. D. Fuhlendorf, and S. L. Winter. 2011.
Aboveground macroinvertebrate diversity and abundance in sand sagebrush
prairie managed with the use of pyric herbivory. Rangel. Ecol. Manag. 64:
394–403.
Engle, D. M., M. K. Fuhlendorf, A. Roper, and D. M. Leslie. 2008.
Invertebrate community response to a shifting mosaic of habitat. Rangel
Ecol. Manag. 61: 55–62.
Fuhlendorf, S. D., and D. M. Engle. 2001. Restoring heterogeneity on range-
lands: Ecosystem management based on evolutionary grazing patterns.
Bioscience 51: 625–632.
Grixti, J. C., L. T. Wong, S. A. Cameron, and C. Favret. 2009. Decline of bum-
ble bees (Bombus) in the North American Midwest. Biol. Conserv. 142:
75–84.
Harrison, S., B. D. Inouye, and H. D. Safford. 2003. Ecological heterogeneity
in the effects of grazing and fire on grassland diversity. Conserv. Biol. 17:
837–845.
Hayes, H. F., and K. D. Holl. 2003. Cattle grazing impacts on annual forbs
and vegetation composition of mesic grasslands in California. Conserv.
Biol. 17: 1694–1702.
Helzer, C. J., and A. A. Steuter. 2006. Preliminary effects of patch-burn graz-
ing on a high diversity prairie restoration along the Central Platte River,
Nebraska. Ecol. Restor. 23: 167–171.
Hendrix, S. D., K. S. Kwaiser, and S. B. Heard. 2010. Bee communities
(Hymenoptera: apoidea) of small Iowa hill prairies are as diverse and rich as
those of large prairie preserves. Biodivers. Conserv. 19: 1699–1709.
Hines, H., and S. D. Hendrix. 2005. Bumble bee (Hymenoptera: Apidae) di-
versity and abundance in tallgrass prairie patches: the effects of local and
landscape features. Environ Entomol. 34: 1477–1484.
Holecheck, J. L., R. D. Pieper, and C. H. Herbel. 2003. Range management
principles and practices. Prentice Hall, Upper Saddle River, NJ.
Hovick, T. J., J. R. Miller, S. J. Dinsmore, D. M. Engle, D. M. Debinski, and S.
D. Fuhlendorf. 2012. Effects of fire and grazing on grasshopper sparrow
nest survival. J. Wildl. Manag. 76: 19–27.
JMPVR . 2012. version 10.0.0. SAS Institute Inc., Cary, NC.
Kirchner, B. N., N. S. Green, D. A. Sergeant, and J. N. Mink. 2011. Responses
of small mammals and vegatation to a prescribed burn in a tallgrass black-
land prairie. Am. Nat. J. 166: 112–125.
Kremen, C., R. L. Bugg, N. Nicola, S. A. Smith, R. W. Thorp, and N. M.
Williams. 2002. Native bees, native plants, and crop pollination in
California. Femontia 30: 41–49.
Kruess, A., and T. Tscharntke. 2001. Grazing intensity and the diversity of
grasshoppers, butterflies, and trap-nesting bees and wasps. Conserv. Biol.
16: 1570–1580.
Kruess, A., and T. Tscharntke. 2002. Contrasting responses of plant and insect
diversity to variation in grazing intensity. Biol. Conserv. 106: 293–302.
Magurran, A. E. 2004. Measuring biological diversity. Blackwell Science,
Oxford, United Kingdom.
McGranahan, D. A., D. M. Engle, B. J. Wilsey, S. D. Fuhlendorf, J. R. Miller,
and D. M. Debinski. 2012. Grazing and an invasive grass confound spatial
pattern of exotic and native grassland plant species richness. Basic Appl.
Ecol. 13: 654–662.
Miller, J. R., L. W. Morton, D. M. Engle, D. M. Debinski, and R. N. Harr.
2012. Nature reserves as catalysts for landscape change. Front Ecol.
Environ. 10: 144–152.
Moeller, D. A. 2005. Pollinator community structure and sources of spatial
variation in plant-pollinator interactions in Clarkia xantiana ssp. xantiana.
Oecologia 142: 28–37.
Moranz, R. A. 2010. The effects of ecological management on tallgrass prairie
butterflies and their nectar sources. Dissertation. Oklahoma State
University, Stillwater, Oklahoma.
Moranz, R. A., D. M. Debinski, D. A. McGranahan, D. M. Engle, and J. R.
Miller. 2012. Untangling the effects of fire, grazing, and land-use legacies on
grassland butterfly communities. Biodivers. Conserv. 2111: 2719–2746.
Moranz, R. A., D.M. Debinski, L.Winkler, J. Trager, D. A.McGranahan, D.M.
Engle, and J. R.Miller. 2013. Effects of grassland management practices on ant
functional groups in central North America. J. Insect Conserv. 17: 699–713.
Moranz, R., S. Fuhlendorf, and D. Engle. 2014. Making sense of a prairie
butterfly paradox: The effects of grazing, time since fire, and sampling
period on regal fritillary abundance. Biol. Conserv. 173: 32–41.
Morse, D. H. 1980. The effect of nectar abundance on foraging patterns of
bumble bees. Ecol. Entomol. 5: 53–59.
Myack, C., and D. Naug. 2009. Energetic stress in the honeybee Apis mellifera
from Nosema ceranae infection. J. Invertebr. Pathol. 100: 185–188.
National Research Council. 2006. Status of pollinators in North America.
The National Academies Press, Washington, DC, p. 322.
Panzer, R., and M. W. Schwartz. 2000. Effects of management burning on
prairie insect species richness within a system of small, highly fragmented
reserves. Biol. Conserv. 96: 363–369.
Pelletier, L., J. N. McNeil. 2003. The effect of food supplementation on repro-
ductive success in bumblebee field colonies. Oikos 103: 688–694.
Pereboom, J. J. M. 2000. The composition of larval food and the significance
of exocrine secretions in the bumblebee Bombus terrestris. Insect Soc. 47:
11–20.
Pillsbury, F. C., J. R. Miller, D. M. Debinski, and D. M. Engle. 2011. Another
tool in the toolbox? Using fire and grazing to promote bird diversity in
highly fragmented landscapes. Ecosphere 28: 1–14.
Platt, W. J., G.W. Evans, andM.M. Davis. 1988. Effects of fire season on flow-
ering of forbs and shrubs in longleaf pine forests. Oecologia 76: 353–363.
Pleasants, J. M. 1981. Bumblebee response to variation in nectar availability.
Ecology 1981: 1648–1661.
Rambo, J. L., and S. H. Faeth. 1999. Effect of vertebrate grazing on plant and
insect community structure. Conserv. Biol. 13: 1047–1054.
Reed, C. C. 1997. Responses of prairie insects and other arthropods to pre-
scription burns. Nat. Areas J. 17: 380–385.
Ricketts, T. H., J. Regetz, I., Steffan-Dewenter, S. A. Cunningham, C.
Kremen, A. Bogdanski, B. Gemmill-Herren, S. S. Greenleaf, A. M. Klein,
et al. 2008. Landscape effects on crop pollination services: are there general
patterns? Ecol. Lett. 11: 499–515.
Rudolph, D. C., C. A. Ely, R. R. Schaefer, J. H. Williamson, and R. E. Thill.
2006. The diana fritillary (Speyeria diana) and great spangled fritillary
(S. cybele): Dependence on fire in the ouachita mountains of Arkansas.
J. Lepidopterists’ Soc. 60: 218–226.
Slagle, M. W., and S. D. Hendrix. 2009. Reproduction of Amorpha canescens
(Fabaceae) and diversity of its bee community in a fragmented landscape.
Comm. Ecol. 161: 813–823.
Steffan-Dewenter, I., A. Kuhn. 2003. Honeybee foraging in differentially
structured landscapes. Proc. R. Soc. Lond. B 270.1515: 569–575.
Sutcliffe, G. H., and R. C. Plowright. 1990. The effects of pollen availability
on development time in the bumble bee Bombus terricola K. (Hymenoptera:
Apidae). Can. J. Zool. 6: 1120–1123.
Swengel, A. B. 1998. Effects of management on butterfly abundance in tall-
grass prairie and pine barrens. Biol. Conserv. 831: 77–89.
Swengel, S., D. Schlicht, F. Olsen, and A. Swengel. 2011. Declines of prairie
butterflies in the midwestern USA. J. Insect Conserv. 15: 327–339.
Toth, A. L., S. Kantarovich, A. F. Meisel, and G. E. Robinson. 2005.
Nutritional status influences socially-regulated foraging ontogeny in honey
bees. J. Exp. Biol. 208: 4641–4649.
vanEngelsdorp, D., and M. D. Meixner. 2010. A historical review of managed
honey bee populations in Europe and the United States and the factors that
may affect them. J. Invert. Pathol. 103: S80–S95.
Vogel, J. A., D. M. Debinski, R. R. Koford, and J. R. Miller. 2007. Butterfly
responses to prairie restoration through fire and grazing. Biol. Conserv. 140:
78–90.
Vogel, J., R. Koford, and D. M. Debinski. 2010. Direct and indirect effects of
time since burn on butterfly abundance in tallgrass prairie. J. Insect
Conserv. 14: 663–677.
von Frisch, K. 1965. Tanzsprache und Orientierung der Bienen. Springer,
Berlin, Heidelberg, New York.
Westphal, C., I. Steffan-Dewenter, and T. Tscharntke. 2009. Mass flowering
oilseed rape improves early colony growth but not sexual reproduction of
bumblebees. J. Appl. Ecol. 46: 187–193.
Winfree, R., R. Aguilar, D. P. Vazquez, G. LeBuhn, and M. A. Aizen. 2009. A
meta-analysis of bees’ responses to anthropogenic disturbance. Ecology 90:
2068–2076.
Winston, M. L. 1987. The Biology of the Honey Bee. Harvard University
Press, Cambridge, MA.