ArticlePDF Available

Abstract and Figures

1.Bee pollinators provide a critical ecosystem service to wild and agricultural plants but are reported to be declining worldwide due to anthropogenic change. Long-term data on bee abundance and diversity are scarce, and the need for additional quantitative sampling using repeatable methods has been emphasized. Recently, monitoring programs have begun using a standardized method that employs a combination of pan traps and sweep netting, resulting in lethal sampling of bees. This standardized method can remove a large number of bees from sites during each sampling day, raising concern that the sampling itself could have a negative effect on bee populations.2.We conducted an experiment to assess whether lethal sampling for bees using pan traps and netting affected bee abundance and diversity when done every two weeks throughout a season and over multiple years. We compared bee abundance, richness, evenness and functional group composition between sites that had been sampled every two weeks from 2009 to 2012 to similar sites not previously sampled.3.We found that the standardized method for sampling bees, with specimens from 132morphospecies, did not affect bee communities in terms of abundance, rarefied richness, evenness, or functional group composition. Thus, our results indicate that the bee communities we sampled are robust to such sampling efforts, despite removing an average of 2,862 bees per season.4.We discuss several explanations for why sampling did not affect bee abundance or community structure, including a density-dependent response to reduced competition for resources.5.These results suggest that bee monitoring programs sampling once every two weeks with pan traps and netting will not affect bee community structure. We urge researchers monitoring bees to utilize standardized protocols so that results can be compared across space and time.This article is protected by copyright. All rights reserved.
Content may be subject to copyright.
The effect of repeated, lethal sampling on wild bee
abundance and diversity
Zachariah J. Gezon
1,2
*,EliS.Wyman
3
, John S. Ascher
4
,DavidW.Inouye
2,5
and
Rebecca E. Irwin
1,2
1
Department of Biological Sciences, Dartmouth College, 78 College St, Hanover, NH 03755, USA;
2
Rocky Mountain Biological
Laboratory, Crested Butte, CO 81224, US A;
3
Division of Invertebrate Zoology, American Museum of Natural History, New
York, NY 1 0024, USA;
4
Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, Singapore
117543, Singapore; and
5
Department of Biology, University of Maryland, College Park, MD 20742, USA
Summary
1. Bee pollinators provide a critical ecosystem service to wild and agricultural plants but are reported to be
declining world-wide due to anthropogenic change. Long-term data on bee abundance and diversity are scarce,
and the need for additional quantitative sampling using repeatable methods has been emphasized. Recently,
monitoring programmes have begun using a standardized method that employs a combination of pan traps and
sweep netting, resulting in lethal sampling of bees. This standardized method can remove a large number of b ees
from sites during each sampling day, raising concern that the sampling itself could have a negative effect on bee
populations.
2. We conducted an experiment to assess whether lethal sampling for bees using pan traps and netting affected
bee abundance and diversity when done every two weeks throughout a season and over multiple years. We com-
pared bee abundance, richness, evenness and functional group composition between sites that had been sampled
every two weeks from 2009 to 2012 to similar sites not previously sampled.
3. We found that the standardized method for sampling bees, with specimens from 132 morphospecies, did not
affect bee communities in terms of abundance, rarefied richness, evenness, or functional group composition.
Thus, our results indicate that the bee communities we sampled are robust to such sampling efforts, despite
removing an average of 2862 bees per season.
4. We discuss several explanations for why sampling did not affect bee abundance or community structure,
including a density-dependent response to reduced competition for resources.
5. These results suggest that bee monitoring programmes sampling once every two weeks with pan traps and net-
ting will not affect bee community structure. We urge researchers monitoring bees to utilize standardized proto-
cols so that results can be compared across space and time.
Key-words: bee bowl, native bees, over-collecting, over-sampling, pan traps, Pollination, pollinator
decline
Introduction
Environmental monitoring programmes are fundamental to
ecological and environmental sciences, and play a crucial role
in informing science-based environmental policy (Lovett et al.
2007). Environmental monitoring programmes are commonly
used to assess environmental change, such as the effects of sub-
urbanization on pollinator communities (Carper et al. 2014),
the effects of habitat restoration on insect food webs (Albrecht
et al. 2007), and the impact of dams on river benthic inverte-
brate communities (Statzner et al. 2001; Chaves-Ulloa, Uman-
a-Villalobos & Springer 2014). Many environmental
monitoring programmes require the collection and killing of
organisms due to the difficulty of identifying specimens in the
field (Reynoldson & Metcalfe-Smith 1992; Hawkins et al.
2000; Schonberg et al. 2004), creating concern that the moni-
toring itself may negatively impact the study taxa or system,
and that changes do cumente d by monit oring prog rammes
may be caused by the monitoring efforts themselves. Monitor-
ing of wild bees is of particular interest due to the pollination
services they provide (Gallai et al. 2009) and the perception
that bees are in decline (Potts et al. 2010), but most bees are
difficult to identify in the field, necessitating collection to make
or confirm species identification (Tepedino & Stanton 1981;
Gibbs et al. 2013). The aim of this study was to determine
whether a commonly employed bee monitoring protocol using
lethal sampling techniques affected bee abundance and diver-
sity when used repeatedly over several years.
Pollination by insects is an essential ecosystem service in
both natural and agricultural systems. Approximately 90% of
flowering species world-wide rely on pollination by insects and
other animals for reproduction (Ollerton, Winfree & Tarrant
*Corresponding author. E-mail: zachariah.j.gezon.gr@dartmouth.edu
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society
Methods in Ecology and Evolution 2015 doi: 10.1111/2041-210X.12375
2011), and pollinators provide approximately $202 billion per
year globally in pollination services to agricultural systems
(Gallai et al. 2009). Despite the ecological and economic
importance of insect pollinators, and of bees in particular, little
is known about their population trends beyond what can be
learned from comparisons to historical records or museum
specimens (Bartomeus et al. 2013; Burkle, Marlin & Knight
2013), prompting the development of standardized protocols
for sampling bees (e.g. LeBuhn et al. 2003; Westphal et al.
2008). Concerns about global pollinator declines along with
the development of standardized protocols have motivated the
collection of long-term bee pollinator data using monitoring
programmes (Committee on the Status of Pollinators in North
America, & National Research Council 2007; Westphal et al.
2008) and the proposal of large-scale, long-term community
efforts (Lebuhn et al. 2013).
One commonly used standardized bee catching protocol
(outlined in Materials and Methods) (e.g. Wilson, Messinger &
Griswold 2009; Neame, Griswold & Elle 2013; Winfree et al.
2014) employs a combination of sweep netting and pan traps
(small plastic b owls pain ted uorescent yellow, fluorescent
blue, or white, filled with soapy water). Pan traps typically
account for a large percentage of bees caught (Wilson, Gris-
wold & Messinger 2008; Popic, Davila & Wardle 2013) and, by
design, kill the sampled bees. Thus, a large number of bees can
be collected each time the protocol is used, and bees are typi-
cally sampled repeatedly in the same site over the course of a
season and sometimes for multiple years (e.g. Wilson, Gris-
wold & Messinger 2008; Kearns & Oliveras 2009; Jha & Van-
dermeer 2010). A concern is that high bee capture rates of
certain taxa, such as primitively eusocial Lasioglossum (Dialic-
tus), obtained using pan traps and netting could affect bee
abundance, diversity, and community composition. Uninten-
tionally affecting wild bee populations would be problematic
because th e me thods could (i) drive long-term patterns in mea-
sured bee abundance, diversity, and community composition,
thus sampling would not accurately reflect how unsampled bee
populations and communities would have responded to their
natural environment, and (ii) negatively impact pollination in,
and near, the study sites. Although concern has been voiced
about the potential effects of over-collecting bees (Shepherd
et al. 2003; Tepedino et al. 2015), little work to date has inves-
tigated the long-term impacts of pan trapping and netting.
The purpose of this study was to assess whether repeated use
of pan trapping and netting had an impact on abundance and
diversity of bee populations and communities. We used two
approaches. First, we compared trends in bee abundance over
5 years of sampling from sites that were sampled every
2 weeks throughout the ight season for bees. We predicted
that if repeated sampling negatively affected bee populations,
bee abundance would decrease both over the course of a sam-
pling season, and from one year to the next. Secondly, we used
an experiment in which we compared bee abundance, species
richness and evenness at sites that had been sampled twice per
week for 4 years to sites that had never been sampled before.
We predicted that if bee communities were affected by repea-
ted pan trapping and netting, bee abundance and diversity
(richness and evenness) would be lower in sites that had been
repeatedly sampled than in newly sampled sites.
Materials and methods
STUDY AREA
We conducted this study in the West Elk Mountains of southwest
Colorado, in Gunnison County near the Rocky Mountain Biological
Laboratory (RMBL). The RMBL (2886 m elevation, 38°57
0
N,
106°59
0
W) is highly seasonal, with a growing season typically from late
April or May until September (CaraDonna, Iler & Inouye 2014). Open
meadow habitats around the RMBL are dominated by numerous
perennial flowering plant species, many of which are visite d by bee p oll-
inators (Forrest, Inouye & Thomson 2010). Pollinator communities
around the RMBL are dominated by native bees (Inouye 1977; Forrest
& Thomson 2011); the European Honey Bee Apis mellifera (Hymenop-
tera: Apidae) is absent in this study system, and other non-native bee
species nearly so.
STUDY SITES
We studied a total of 26 sites in three river valleys: East River, Slate
River, and Washington Gulch (Appendix S1 and Fig. S1). To test
whether bee sampling affected intra- and inter-annual variation in bee
abundance, we used nine sites, all located in the East River valley
(Appendix S1). For five consecutive years (20092013), these sites were
sampled every two weeks throughout the flowering season. Each site
was classified as either wet or dry (3 sites) meadow, and wet meadows
were further divided into wet-Veratrum (3 sites) and wet-Salix (3 sites)
based on dominant vegetation (Veratrum tenuipetalum and shrubby
Salix spp ., respectively). Wet and d ry meadow are two dominant habi-
tat types around the RMBL (Langenheim 1962). To test whether sites
that were sampled repeatedly every two weeks for three years versus
only sampled once differed in bee abundance and diversity in 2012, we
used 23 sites. Seventeen were sampled only once (11 dry meadow and
six wet-Veratrum sites), and were located in the East, Slate, and Wash-
ington Gulch river valleys (Appendix S1). We compared these ‘single-
sample’ sites to sites of similar habitats that had been sampled every
2 weeks from 2009 to 2012 (the three wet-Veratrum and three dry mea-
dow sites described above, Appendix S1). D uring the summer of 2012,
each time we sampled a repeat-sample site, we would sample a compa-
rable single-sample site within two days. We then assigned a numeric
‘sample week’ to the single-sample and repeat-sample sites that were
blocked in time, for a total of seven sample weeks.
BEE SAMPLING
Each site consisted of two, fixed 45 m transects at approximately right
angles. Pan traps were made from 9 cm diameter plastic bowls (Fisher-
brand
Ò
; Hexagonal Polystyrene Weighing Dishes, Pittsburgh, PA,
USA) painted fluorescent yellow or fluorescent blue using Krylon
Ò
Fluorescent Paint (colour numbers 3104 and 3109, respectively), or left
unpainted (white). On each sample day, pan traps filled with soapy
water (water with a small amount of Dawn
Ò
Ultra Dishwashing soap,
original scent) were placed every 3 m alon g each transect with pan-trap
colours assigned at random (as in LeBuhn et al. 2003). Pan traps were
left in place from ~0800 to 1700, the peak hours of pollinator activity
around the RMBL, after which the contents were strained and stored
in 75% EtOH until processing. All bees were washed in water, dried
with a hair drier, pinned and labelled as in LeBuhn et al. (2003).
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
2 Z. J. Gezon et al.
Pan traps typically catch many smaller bees but fail to consistently
capture larger bodied, stronger flying bees including some Megachili-
dae and Apidae (Cane, Minckley & Kervin 2000). Thus, in conjunction
with the pan traps, we systematically walked each transect, netting for
bees for one hour each in the morning and afternoon when pan traps
were in place (as outlined in LeBuhn et al. 2003). Netted bees that
could not be identified to species in the field were c ollected, pinned, and
labelled. We could identify most bumblebees (Bombus)tospeciesinthe
field based on pile colours, with reference to the RMBL insect collec-
tion. Ide ntified bumblebees were marked with a dot of indelible ink on
their thorax to avoid recounting released individuals.
Pinned and labelled bees from the pan traps and netting were identi-
fied to species or morphospecies, although we were unable to identify
some cryptic taxa beyond genus or subgenus, notably Lasioglossum
(Dialictus)andL. (Evylaeus sensu lato).
To control for potential differences in biotic and abiotic conditions
between sites sampled repeatedly versus once, we collected the follow-
ing covariates each time we sampled for bees. First, along the transects
at each site, we measured floral abundance using three, fixed, 4 m
2
flower census plots (six per site). On the same days that the sites were
sampled for bees, flower censuses were conducted within all plots by
recording all flowering plant species, number of flowers per plant for
up to ten plants per species, and number of plants per plot. Secondly,
we recorded daily maximum temperatures during each sampling period
from a c entrally located weather station near the RMBL (38
°
57
0
47N,
106
°
59
0
36W, 2945 m elevation). Temperature can vary dramatically
at these subalpine sites, with strong effects on insect flight and activity
(Corbet et al. 1993).
ANALYSES
We performed separate analyses for bees that were captured and killed
(hereafter referred to as ‘vouchered’), bees that were netted, identified
in the field and released alive (hereafterreferredtoas‘released),aswell
as with all bee data combined (hereafter referred to as ‘combined’).
Only vouchered specimens are included in the analyses described below
except where otherwise indicated. Due to extreme drought conditions
in 2012, abundance of Bombus workers was extremely low at all sites
(both those that were sampled repeatedlyaswellasthoseonlysampled
once); thus, an analysis of only the re leased bees was not possible. For
eachsamplingbout,thebeecapturerate(aproxyforabundance)was
calculated as bees caught per hour (bees/hour). For vouchered bees, we
averaged the capture rates for bees caught in pan traps and bees caught
by net. Floral density (flowers m
2
) was calculated by multiplying the
average num ber of flowers per plant for each species by the number of
plants per plot, summing across species, and dividing by four (each plot
was 4 m
2
). Plot means were then averaged at each site for each sample
bout to calculate site mean floral density on each sampling day. Bee
species richness was rarefied to the lowest common abundance across
the combined data using the vegan package in R version 3.0.3 (R
Development Core Team 2011; Oksanen et al. 2012), and evenness
was calculated as E
var
(Smith & Wilson 1996). E
var
ranges from 0 to 1,
with 0 representing minimum evenness and 1 maximum evenness
(Smith & Wilson 1996).
We also wanted to determine whether community composition of
particular bee functional groups (Winfree, Griswold & Kremen 2007;
Williams et al. 2010) was affected by pan trapping. We categorized
the nesting habits of eac h species (soil, wood, pithy stems, constructed
cells, cavity or hive), diet (pollen generalists vs. pollen specialists), and
sociality (eusocial vs. not eusocial) based on information compiled
from published literature (Mitchell 1960; Hurd 1979; Cane, Griswold
& Parker 2007; Michener 2007) and from extensive field experience by
J.A. and E.W. We used Carper et al. (2014) as the basis for our nest-
ing substrate designations, with the addition of the category ‘con-
structed cells’ for bees that construct cells on exposed surfaces using
resins, masticated leaves, and other materials. We differentiated
between bees that nest in existing cavities in wood (‘cavity’) and those
that excavate nests in this substrate (‘wood’), although the results were
not qualitatively affected when ‘cavity’ and ‘wood’ were combined
(analyses not shown). We classified all non-parasitic Bombus as ‘hive’
nesters due to the heterogeneity of nesting substrates used by Bombus,
even within species. Parasitic bees were included in the nesting cate-
gory of their host. The category diet’ refers to pollen specialists versus
pollen generalists, with ‘specialist’ referring to known oligolectic spe-
cies,aswellastaxaforwhichavailable evidence suggests oligolecty or
mesolecty. ‘Sociality’ was broken into two categories: ‘eusocial’ and
‘not eusocial’, with ‘eusocial’ containing only non-parasitic Bombus
spp. Species that could potentially be eusocial based on their phyloge-
netic placement, such as Halictus and Lasioglossum (Dialictus), were
considered ‘not eusocial’ because eusociality is unlikely to be realized
in high-elevation habitats with short growing seasons (Kocher et al.
2014), such as subalpine Colorado. Nonetheless, we also conducted a
separate analysis of sociality, with Bombus, Halictus and Lasioglossum
(Dialictus) treated as ‘eusocial’ given that in lower elevation areas
not sampled in this study, Halictus and/or Lasioglossum (Dialictus)
can be eusocial (Kocher et al. 2014). Parasitic bees were excluded
from analyses of diet and sociality. We also categorized bee species by
size using intertegular (IT) span to estimate dry body mass (Cane
1987; Greenleaf et al. 2007). We measured IT span using a stereo-
scope and calibrated ocular micrometer for up to five randomly
selected female bees from each morphospecies, using only worker bees
for social species. We classified species as small (<4mg),medium
(416 mg), or large (>16 mg) (Winfree, Griswold & Kremen 2007).
We performed separate analyses for bee abundance (estimated as the
bee catch rate), rarefied richness, evenness, and functional groups. We
also calculated mean expected foraging distance for the bees using
overall mean IT span and the statistical relationship between body
size and foraging distance presented by Greenleaf et al. (2007).
To test whether bee abundance varied over the course of each sam-
pling season and from year to year, we used generalized linear mixed
models. We used AICc values to rank and compare multiple
apriori
hypotheses and pick the model best supporte d by th e data. The candi-
date models included meadow type (dry, wet-Veratrum,orwet-Salix),
mean floral density, and daily maximum temperature as factors. Site,
year and sampling week were included as random effects. Sampling
week was defined as the number of times a site was sampled within a
season, which allowed for differences in timing of snowmelt (and thus
accesstothesiteforsampling)among sites and years. For example,
sampling at higher elevation sites began slightly later than lower eleva-
tion sites due to later snowmelt and phenology, but the first time each
site was sampled was lumped into a single phen ologically similar sam-
ple week. Likewise, years with early snowmelt and phenology (e.g.
2012) had more sampling weeks than years with late snowmelt and phe-
nology (e.g. 2011) due to a longer flight season. We log-transformed
bee capture rates to fulfil the assumption of normality (here and below).
We used the lmer function in th e lme4 package to construct the models
and the aictab function in the AICcmodavg package to rank and com-
pare models using R version 3.0.3 (R Development Core Team 2011;
Mazerolle 2013; Bates et al. 2014). We used type III sums of squares
and Satterthwaite approximation of degrees of freedom to calculate F
and P values.
To test for differences in bee abundance, rarefied richness, and even-
ness between single-sample sites and repeat-sample sites, we used nested
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
Effect of lethal sampling on bee populations 3
ANCOVAs with treatment (single or repeated sampling), block (week),
and river valley as factors, and mean floral density and daily maximum
temperature as covariates. We included block as a random effect to
control for the week the samples were collected, and we tested treat-
ment over the block term (here and below). To test whether the effect of
repeated sampling varied with bee functional group, we used general
linear mixed models with functional group, treatment (single or
repeated sampling), block (week), and river va lley as factors, and mean
floral d ensity and daily maximum temperature as covariates. We
included site as a rando m effect in the functional group analysis to con-
trol for variation among sites. Four separate analyses were performed
for eac h functional group: nesting habits, diet, sociality, and body size.
We used JMP 11.0 (SAS Institute 2013) to perform the nested
ANCOVAs
and functional group analyses. To visualize potential differences in
community composition between single and repeat-sampled sites, we
used non-metric multidimensional scaling (NMDS), using the Bray
Curtis dissimilarity index, using the vegan package (Oksanen et al.
2012) in R version 3.0.3 (R Development Core Team 2011). Stress val-
ues from models with two to six dimensions were compared to deter-
mine th e most appropriate summary of the relationship between
sampling and community composition.
Results
DATA OVERVIEW
Across five years and 26 sites, we vouchered 14 312 pan-
trapped bees, netted-and-vouchered 995 bees, and netted-
and-released 2489 bees. Collectively, these bees belonged to
five families, 28 genera, and 132 species/morphospecies
including 26 county records and three state records (Appen-
dix S2), with a mean IT span of 139 mm and mean expected
foraging range of 66 m (Greenleaf et al. 2007). Approxi-
mately 28% of the bees caught were male, and 40% of the
bees caught were Lasioglossum (Dialictus)sweatbees.For
bees identified to species, the two most common species were
the sweat bees Halictus virgatellus and Halictus rubicundus,
comprising 1131% and 2 52% respectively of all bees
vouchered. In addition, we captured 35 species represented by
a single individual. In 2012, we collected 100% of the genera
and 98% of the morphospecies that we would eventually
catch after just 62 of the 95 total sampling events. A rarefac-
tion curve of the 2012 sampling data (Fig. S2) revealed a satu-
rating function, suggesting that our sampling effort
adequately characterized common bee taxa but that we would
likely continue to capture new, rarer species with additional
sampling.
INTER- AND INTRA-ANNUAL VARIATION
Although we predicted that repeated sampling of sites would
result in a decrease in bee catch rate both among years and
within seasons, we found no evidence for either pattern.
Among years, total bee catch rates (vouchered and released
combined, divided by total sampling effort) ranged from
633 bees per hour in 2011 to 814beesperhourin2010
(Fig. 1). Catch rates for vouchered bees only were consistent
across years, except for 2012, when the catch rate nearly
doubled. Moreover, when we examined intra-annual varia-
tion in vouchered bee catch rates, we found no general pat-
tern within years (assessed visually from Fig. 2, S3 and S4).
The best-fit model for analysing inter- and intra-annual vari-
ation in bee abundance was supported by 88% of the
Akaike weights and included significant interactions between
flower density and maximum daily temperature, with site as
a random effect. The model also allowed the intercepts (but
not slopes) to vary among years and among sample weeks
within years (nested random effects) (Appendix S3). In addi-
tion, we found significant main effects of both flower density
and maximum daily temperature on vouchered bee catch
rate (F
1,26473
= 983 , P < 0002 and F
1, 27697
= 1227,
P < 0001, respectively), with bee catch rate increasing with
increasing flower density and temperature (b 1SE:
048 015 and 008 002, respectively). We also found a
Fig. 1. Interannual variation in mean bee
catch rates ( 1 SE)forbeesthatwerecaught
in pan traps and vouchered (black bars),
caught in nets and vouchered (white bars),
andcaught,identiedinthefieldandreleased
(gray bars). Only Bombus spp. were identified
in the field and released. Bee catch rates varied
among years with no consistent pattern over
time.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
4 Z. J. Gezon et al.
significant interaction between flower density and maxi-
mum daily temperature on vouchered bee catch rate (b =
002 0007, F
1, 26885
= 945, P = 00023).
REPEATED VERSUS SINGLE-SAMPLED SITES
The catch rate of vouchered bees was 3128% higher on aver-
age in repeat-sampled sites versus single-sampled sites,
although the difference was not statistically significant
(Table 1, Fig. 3A). We also found no differences in rarefied
richness or evenness (E
var
) between repe at-samp led and single-
sampled sites (Table 1, Fig. 3B and 3C, respectively). River
valley had a significant effect on rarefied species richness, with
the East River valley having higher rarefied richness (Least
Squares Means SE: 568 022 species) than Slate River
valley (442 058 species) and Washington Gulch
(364 056 species). Bee catch rate was 5954% higher in
wet-Veratrum sites than dry meadow sites (1264 503 and
792 309, respectively). We also found a weak, positive
relationship between rarefied richness and evenness and maxi-
mum daily temperature, and a weak, positive relationship
between catch rate of vouchered bees and flower density
(Table 1). Results were qualitatively similar for combined bee
data (Fig. S5 and Appendix S4).
We also found no significant main effect of treatment
(repeated versus single-sample) when we subdivided the bees
by functional group (nesting habits, diet, sociality, and body
size) (Table 2, Fig. 4). Likewise, we found no significant inter-
action between treatment and functional group for nesting
habits and body size (Table 2, Fig. 4a,d). When only Bombus
(and not Halictus and Lasioglossum (Dialictus)) were treated as
eusocial, we found no significant interaction between treat-
ment and sociality (Table 2, Fig 4c; see Appendix S5, Fig. S6
for analysis of these taxa categorized as eusocial). We found a
significant interaction between treatment and diet (Table 2,
Fig. 4b), with catch rates of pollen specialists being 4388%
higher, but catch rates of pollen generalists being 795% lower,
in repeated- than single-sample sites. Bee catch rate varied by
Fig. 2. Bee catch rates (mean bees caught per
hour) for vouchered bees (collected using pan
traps and netting) varied from year to year,
with no con sistent pattern related to the num-
ber of times the sites were sampled. Data rep-
resent mean ( 1 SE) catch rates across all
sites after repeated sampling within the season.
Years with earlier phenology (e.g., 2012) had
more sampling bouts due to the longer flying
season.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
Effect of lethal sampling on bee populations 5
functional group classification (Table 2). Soil-nesting bees
comprised 6663% of all vouchered bees, and were between
406 and 3542-times more common than bees of any other
nesting habits (Fig. 4a). The majority of vouchered specimens
were pollen generalists (84 81%, Fig. 4b) and not eusocial
(9562%, assuming our montane Halictus and Lasioglossum
(Dialictus) were not eusocial, Fig. 4c). Small- (4891%) and
medium-sized (3875%) bees were both more common than
large bees (12 33%, Fig. 4d).
NMDS indicated overlap in bee community composition
between repeated versus single-sampled sites (Fig. 5a),
among river valleys (Fig. 5b), and between wet versus dry
habitats (Fig. 5c). Community composition changed dramat-
ically over the course of the summer, as indicated by the dis-
tinct clusters of points when separated by sample week
(Fig. 5d). The overall stress of the two-dimensional NMDS
was 018, indicating fair confidence in the ordination dis-
tances. Models with higher dimensionality resulted in lower
stress (e.g. three-dimensional solution, stress = 0126), but
the results were qualitatively similar (Fig. S7ad) to the two-
dimensional solution.
Discussion
Ourresultssuggestthatpantrappingforbeesdoesnotaect
bee catch rate or diversity, even after sampling the same sites
every other week during the flight season for three consecutive
years. This lack of response was consistent across all bee func-
tional groups with the exception of diet. Although we caught
more pollen specialists in repeated-sample sites than in single-
sample sites (and the opposite pattern for pollen generalists),
the magnitude of effect associated with diet was relatively small
and did not represent a large change in catch rate between
treatments (Fig. 4b). Taken together, these results suggest that
the bee populations sampled in this study are robust to
repeated, lethal sampling twice monthly for multiple, consecu-
tive years.
Several non-mutually exclusive mechanisms could explain a
lack of sampling effect, four of which we explain here. The first
two mechanisms reflect underlying ecological mechanisms
relatedtopopulationdynamics,andthesecondtwoarebased
on species life histories. First, compensatory mortality and
natality could explain the lack of sampling effect. Bees often
compete for limiting resources, including floral food resources
(Minckley et al. 1994; Roulston & Goodell 2011) and nesting
Table 1. Nested ANCOVAs testing the effect of treatment (repeated vs.
single sampling) on (A) vouchered bee catch rate [log(bees/hour)], (B)
rarefied richness and (C) evenness (measured as E
var
). Type refers to
the habitat type (dry or wet-Veratrum), valley is the river valley sites
were located in (East River, Slate River, or Washington Gulch), and
max temp is maximum temperature
DF F Ratio P
(A) Catch rate
Treatment 1, 893 027 0619
Type 1, 2141 6 31 0020
Valley 2 , 979 227 015 5
Flower density 1, 2146 353 0074
Max temp 1, 1725 138 0255
(B) Rarefied richness
Treatment 1, 822 003 0868
Type 1, 2643 0 57 0457
Valley 2, 1434 534 0019
Flower density 1, 2651 039 0538
Max temp 1, 1576 411 0060
(C) Evenness
Treatment 1, 643 016 0698
Type 1, 2432 1 11 0325
Valley 2, 1094 3
60 0063
Flower density 1, 2434 014 0713
Max temp 1, 1260 380 0074
(a)
(b)
(c)
Fig. 3. Least Square Means (LSM) from ANCOVAs of vouchered bees
betweensinglesampledandrepeatedsampledsites.Wefoundnosig-
nificant difference in (a) bee ca tch rate (bees per hour), (b) rarefied rich-
ness, or (c) evenness between single and repeated sampling sites.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
6 Z. J. Gezon et al.
sites (e.g. Hubbell & Johnson 1977; Potts & Willmer 1997; Stef-
fan-Dewenter & Schiele 2008). Reducing bee populations
through le thal sampling may re duce competition for these
resources and therefore increase the population growth rate
via density-dependent processes. Compensation is a funda-
mental principle of wildlife management (Bartmann, White &
Carpenter 1992), and although bee population dynamics are
not well understood (Steffan-Dewenter & Schiele 2008), the
same principles that allow for sustainable harvesting of fish
(Worm et al. 2009) and hunting of wildlife (Boyce, Sinclair &
White 1999) could apply to ‘harvesting’ of insects in environ-
mental monitoring programmes. Secondly, lethal sampling
may have no effect on subsequent bee population size if sam-
pling simply removed individuals from the population that
would not have had an opportunity to reproduce anyway. If
bees are controlled by bottom-up rather than top-down forces
(e.g. Steffan-Dewenter, Potts & Packer 2005), removing adult
bees from the population via lethal sampling may not reduce
population size as long as enough bees remain to utilize all lim-
iting resources. Relative to the first mechanism proposed, here
we would not expect to find a change in bee population g rowth
rate following lethal sampling.
Thirdly, the reproductive habits of bees may lessen the
effects of sampling. Solitary bees construct and mass provision
brood cells as they forage (e.g. Eickwort et al. 1996); thus,
many individual bees may be sampled after they have mated
and carried out some reproduction, which could mitigate the
effect of sampling among years. Unlike many vertebrate taxa,
solitary bees do not require ongoing parental care after the
female provisions and seals larval brood cells (sealing the
entrance of a completed nest may or may not be essential for
brood survival). Thus, the reproductive output of a vouchered
specimen may be reduced but not eliminated. Moreover, leth-
ally sampling some workers from eusocial species may have
minimal effect on the colony level production of males and
daughter queens (e.g. Harbo 1986). Finally, if males are
removed from the population after mating, or females are not
mate-limited, then sampling males may also have no effect on
univoltine bee populations and communitie s in subsequent
years.
The fourth mechanism we outline concerns the short flight
season of most solitary bees, and thus high species turnover
during a typical summer. In cases where bees have short flight
seasons, sampling every two weeks may not result in within-
season declines in bee abundance if new species or populations
are being sampled during each sampling event. For example,
Halictus rubicundus and other typically social halictine bees
may have solitary populations in high-elevation regions of the
Rocky Mountains due to short flight season, with individual
bees only foraging for approximately two weeks (Eickwort
et al. 1996). Our data support observations of short flight peri-
ods by Eickwort et al. (1996) in that detection of individual
species within our sites spanned 1188 082 days on average
( SE). Thus, in systems with high species (and population)
turnover, a biweekly sampling scheme may allow ample time
for brood provisioning between sampling events.
Although our study specifically tested whether lethal sam-
pling negatively affected bee populations, our results may be
generalizable to other entomological systems. For example,
Hunter et al. (2014) found that 90% of Finnish moth popula-
tions were stable or increasing during 32 years of repeatedly
sampling 11 sites using light traps. Unfortunately, experimen-
tal data on the effects of sampling arthropods are extremely
limited, and most accounts of negative effects of sampling only
address the over-collection of rare and valuable specimens for
the black market, such as alpine Lepidoptera in Asia (Hama,
Ishii & Sibatani 1989). The dearth of information on the effects
of sampling insects is unfortunate given the concerns about
specimen collection in general (Wagner 1991; Minteer, Collins
& Puschendorf 2014b; Minteer et al. 2014a) and of bees in par-
ticular (Tepedino et al. 2015), and the utility of insect collec-
tions for ecological research (e.g. Benton et al. 2002; Novotny
et al. 2007).
Catch rates for both vouchered and released bees varied dra-
matically among years (Fig. 2 and S3, respectively), and there
was no consistent decrease in vouchered bee catch rates
from one year to the next as we would expect if lethal samp ling
Table 2. Nested ANCOVAs testing the effect of treatment (repeated vs.
single sampling) on vouchered bee catch rate [log(bees/hour+1)] by
functional group. Type refers to habitat type (dry or wet-Veratrum
meadow). Separate analyses were performed for each of four functional
groups: (A) nesting substrate, (B) diet, (C) sociality and (D) body size.
Body size was treated as a three-level categorical variable (small, med-
ium and large)
DF F Ratio P
(A) Nesting substrate
Type 1, 12 9618 0 027
Treatment 1, 844 0 15 0711
Valley 2 , 987 1 29 0318
Flower density 1, 3553 2 79 0104
Max temp 1, 1547 0 37 0552
Nesting 5, 191310158 <00001
Nesting*Trmt 5, 1913047 0801
(B) Diet
Type 1, 13 32 5 19 0040
Treatment 1, 877 0 31 0593
Valley 2 , 992 2 86 0105
Flower density 1, 3423 2 30 0138
Max temp 1, 1556 1 83 0196
Diet 1, 49 94 18917 <00001
Diet*Trmt 1, 49 94 6 11 0017
(C) Sociality
Type 1, 1232 3 78 0075
Treatment 1, 7513 0 43 0531
Valley 2 , 988 2 75 0113
Flower density 1, 3184 1 17 0287
Max temp 1, 1301 1 46 0248
Sociality 1, 5186 32625 <00001
Sociality*Trmt 1, 51 86 0 95 0335
(D) Body size
Type 1, 13 67 3 53 0082
Treatment 1, 919 0 18 0678
Valley 2 , 989 2 38 0143
Flower density 1, 3760 3 05 0089
Max temp 1, 1818 0 77 0
393
Size 2, 83 57 53 72 <00001
Size*Trmt 2, 8357 1 49 0231
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
Effect of lethal sampling on bee populations 7
negatively impacted bee populations. Rather, the interannual
variation in bee catch rates was likely related to seasonal differ-
ences among years. For example, high snowfall and late phe-
nology in 2011 resulted in extremely high floral resources, and
may have caused the solitary bee populations to increase,
resulting in high vouchered bee capture rates in the subsequent
year 2012. Likewise, extreme drought and freezing events (Ino-
uye 2008) in 2012 resulted in extremely low floral resources
during that summer, which may have caused the near absence
of Bombus spp. workers. However, the capture rates of both
vouchered and released bees returned to more average num-
bers in 2013, which was a more average year in terms of snow-
pack, phenology and oral abundance (Z. Gezon, personal
observation). Our results reflect the extreme variability and
unpredictability of floral resources observed in many systems,
along with the corresponding effects on pollinator abundance
and plantpollinator interactions (Tepedino & Stanton 1980;
Alarc
on, Waser & Ollerton 2008; Olesen et al. 2008).
Five caveats are important to the interpretation of this
study. First, our results apply to sampling bees every two
weeks and cannot be used to assess whether more frequent
sampling, for example weekly or even daily, may negatively
affect bee populations or communities. Secondly, the distances
between the repeated-sample sites (min = 125 m,
max = 2400 m) were large compared to the mean expected
foraging distance of the bees we sampled (66 m), and thus the
maximum area that we could theoretically depopulate through
sampling. Smaller distances between sampling sites in relation
to bee foraging areas could affect the ability of bees to recolo-
nize sampled areas and could therefore affect bee populations
in some areas. Thirdly, the ecosystem investigated in this study
(subalpine Rocky Mountain West) is highly seasonal with a
short ight-season; thus, there is high species turnover when
sampling once every two weeks. How sampling affects bee pop-
ulations and communities in habitats with less seasonal turn-
over requires further investigation. Fourthly, we did not test
for taxon-specific declines in abundance over sampling sea-
sons. Looking at within-site patterns would be most appropri-
ate for determining whether sampling affects individual
species, but the probability of catching the same taxa within a
site over multiple sampling periodswaslowevenfortheabun-
dant taxa because of the infrequency of sampling compared to
the short flight season of most species within our sites. Fifthly,
we only explored whether repeated sampling affected bee catch
(a) (b)
(c) (d)
Fig. 4. Least Square Means ( LSM) of vouc hered bee catch rates for each of four functional groups, (a) nesting habits, (b) diet, (c) sociality, and (d)
body size. In the sociality analysis, only Bombus were categorized as eusocial. The only functional group classification that interacted significantly
with treatment (single versus repeated sampling) was diet. Note the different scale on the y-axes.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
8 Z. J. Gezon et al.
rate and community structure and did not assess its impact on
floral visitation and plant reproduction. Bees are critical in per-
forming pollination services, and complete removal of even a
single species of bee can have impacts on plantpollinator
interactions and plant reproduction (Brosi & Briggs 2013), so
future studies should investigate whether intensive sampling
results in pollen limitation of melittophilous plants.
Given the economic importance of bee pollinators and the
mounting threat of anthropogenic disturbances such as cli-
mate and land-use change (Klein et al. 2007; Potts et al.
2010; Ollerton, Winfree & Tarrant 2011), bee inventorying
and monitoring programmes are critically important for doc-
umenting the biodiversity of ecosystem service providers.
Long-term and repeated sampling is needed to monitor bee
populations across space and time (Committee on the Status
of Pollinators in North America, & National Research Coun-
cil 2007; Winfree 2010; Lebuhn et al. 2013), but implement-
ing and funding such a programme remains a challenge due,
in part, to divergent views on the relevance and statistical
power of the proposed approaches, and the ability to over-
come the taxonomic impediments (Lebuhn et al. 2013; Tepe-
dino et al. 2015; and response by Lebuhn et al. 2015). The
very effectiveness (lethality) of proposed trapping measures
has also been cited as a concern, but our research provides
compelling evidence that lethal sampling of bees using pan
traps and netting every two weeks does not negatively affect
bee populations in terms of abundance, species richness,
evenness and functional group composition. While our study
does not directly address the mechanisms that make bee pop-
ulations robust to lethal sampling, we suggest several plausi-
ble explanations that draw from well-known ecological
theories and species life histories. However, the generality of
our results remains to be tested for bee populations and com-
munities in other regions. Nonetheless, the lack of environ-
mental impact of bee sampling we document is encouraging,
suggesting that pan trapping and netting bi-monthly may be
a sustainable method to inventory bee diversity and to docu-
ment long-term trends in bee populations and communities.
Acknowledgements
We would like to thank M. G. Rightmyer for identification of Osmia spp.; T. L.
Griswold for identification of Stelis spp.; R. Brennan, R. Polanco, A. Slominski,
S. Turner and J. Welch for field assistance; S. Sprott for making the map of the
field sites; and R. Chaves-Ulloa, members of the Irwin Lab and two anonymous
reviewers for providing valuable feedback on the manuscript. The RMBL and
the Gunnison National Forest provided access to field sites. This work was sup-
ported by grants from the National Science Foundation (DEB-0841862 and
DEB-0922080). Any opinions, findings and conclusions or recommendations
expressed in this material are those of the authors and do not necessarily reflect
the views of the National Science Foundation.
Data accessibility
All supplemental figures and appendices have been uploaded as supporting infor-
mation. Specimen-related data are available at the RMBL website www.rmbl.org
and will be archived at the American Museum of Natural History, which deposits
their data in the Arthropod Easy Capture website and shares it with biodiversity
portals including www.discoverlife.org, and, in the near future, with idigbio.
All other data related to the experiment are available from the Dryad Digital
Repository (Gezon et al. 2015).
References
Alarc
on, R., Waser, N.M. & Ollerton, J. (2008) Year-to-year variation in the
topology of a plant-pollinator interaction network. Oikos, 117, 17961807.
Albrecht, M., Duelli, P., Schmid, B. & Muller, C.B. (2007) Interaction diversity
within quantified insect food webs in restored and adjacent intensively man-
aged meadows. Journal of Animal Ecology, 76, 10151025.
(a) (b)
(c) (d)
Fig. 5. Two-dimensional non-metric multidi-
mensional scaling (NMDS) ordination plots
indicating overlap in bee communities when
separated by (a) treatment (S = single-sam-
pled sites, R = repeat-sampled sites), (b) river
valley (E = East River, S = Slate River,
WG = Washington Gulch), (c) habitat type
(D = dry meadow, W = wet-Veratrum), and
(d) sample week. Abbreviations are located at
the grand mean for each group, and the perim-
eter and area of each group is outlined and
shaded. Note that only vouchered bees were
includedinthisanalysis.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
Effect of lethal sampling on bee populations 9
Bartmann, R.M., Whit e, G.C. & Carpenter, L.H. (1992) Compensatory mortal-
ity in a Colorado mule deer population. Wildlife Monographs, 121,339.
Bartomeus, I., Ascher, J.S., Gibbs, J., Danforth, B.N., Wagner, D.L., Hedtke,
S.M. & Wi nfree, R. (2013) Historical changes in northeastern US bee pollina-
tors related to shared ecological traits. Proceedings of the National Academy of
Sciences of the United States of America, 110, 46564660.
Bates, D., Maechler, M., Bolker, B. & Walker, S. (2014) lme4:Linear mixed-
effects models using Eigen and S4. R package version 1.1-5.
Benton, T.G., Bryant, D.M., Cole, L. & Crick, H.Q.P. (2002) Linking agricul-
tural practice to insect and bird populations: a historical study over three dec-
ades. Journal of Applied Ecology, 39,673687.
Boyce, M.S., Sinclair, A.R.E. & White, G.C. (1999) Seasonal compensation of
predation and harvesting. Oikos, 87, 419426.
Brosi, B.J. & Briggs, H.M. (2013) Single pollinator species losses reduce floral
fidelity and plant reproductive function. Proceedings of the National Academy
of Sciences of the United States of America, 110, 1304413048.
Burkle, L.A., Marlin, J.C. & Knight, T.M. (2013) Plant-pollinator interactions
over 120 years: loss of species, co-occurrence, and function. Science, 339,
16111 615.
Cane, J.H. (1987) Estimation of bee size using intertegular span (Apoidea). Jo ur-
nal of the Kansas Entomological Society, 60, 145.
Cane, J.H., Griswold, T. & Parker, F.D. (2007) Substrates and materials used for
nesting by North American Osmia bees (Hymenoptera: Apiformes: Megachili-
dae). Annals of the Entomological Society of America, 100, 350358.
Cane, J.H., Minckley, R.L. & Kervin, L.J. (2000) Sampling bees (Hymenoptera:
Apiformes) for pollinator community studies: pitfalls of pan-trapping. Journal
of the Kansas Entomological Society, 73, 225231.
CaraDonna, P.J., Iler, A.M. & Inouye, D.W. (2014) Shifts in flowering phenology
reshape a subalpine plant community. Proceedings of the National Academy of
Sciences of the United States of America, 111, 49164921.
Carper,A.L.,Adler,L.S.,Warren,P.S.&Irwin,R.E.(2014)Eectsofsub-
urbanization on forest bee communities. Environmental Entomology, 43,
253262.
Chaves-Ulloa, R., Umana-Villalobos, G. & Springer, M. (2014) Downstream
effects of hydropower production on aquatic macroinvertebrate assemblages
intworiversinCostaRica.Revista de biologia tropical, 62
(Suppl 2), 179201.
Committee on the Status of Pollinators in North America, and National
Research Council. (2007) Status of Pollinators in North America. National
Academies Press, Washington, DC.
Corbet,S.A.,Fussell,M.,Ake,R.,Fraser,A.,Gunson,C.,Savage,A.&Smith,
K. (1993) Temperature and the pollinating activity of social bees. Ecological
Entomology, 18,1730.
Eickwort, G.C., Eickwort, J.M., Gordon, J. & Eickwort, M.A. (1996) Solitary
behavior in a high altitude population of the social sweat bee Halictus rubicun-
dus (Hymenoptera: Halictidae). Behavioral Ecology and Sociobiology, 38,227
233.
Forrest, J., Inouye, D.W. & Thomson, J.D. (2010) Flowering phenology in subal-
pine meadows: does climate variation influence community co-flowering pat-
terns? Ecology, 91,431440.
Forrest, J.R.K. & Thomson, J.D. (2011) An examination of synchrony between
insect emergence and flowering in Rocky Mountain meadows. Ecological
Monographs, 81, 469491.
Gallai, N., Salles, J.M., Settele, J. & Vaissiere, B.E. (2009) Economic valuation of
the vulnerability of world agriculture confronted with pollinator decline. Eco-
logical Economics, 68, 810821.
Gezon, Z.J., Wyman, E.S., Ascher, J.S., Inouye, D.W. & Irwin, R.E. (2015) Data
from: The effect of repeated, lethal sampling on wild bee abundance and diver-
sity. Dryad Digital Repository, doi:10.5061/dryad.5385j.
Gibbs, J., Packer, L., Dumesh, S. & Danforth, B.N. (2013) Revision and reclassi-
fication of Lasioglossum (Evylaeus), L. (Hemihalictus) and L. (Sphecodoga-
stra) in eastern North America (Hymenoptera: Apoidea: Halictidae). Zootaxa,
3672,1117.
Greenleaf, S.S., Williams, N.M., Winfree, R. & Kremen, C. (2007) Bee foraging
ranges and their relationship to body size. Oecologia, 153, 589596.
Hama, E., I shii, M. & Sibatani, A. (1989) Nipponsan tyoorui no suiboo to hogo
(Decline and conservation of butterflies in Japan), Part 1. Nippon Rinsi Gak-
kai (Lepidopterological Society of Japan).
Harbo, J.R. (1986) Effect of population-size on brood production, worker sur-
vival and honey gain in colonies of honeybees. Journal of Apicultural Research,
25,2229.
Hawkins, C.P., Norris, R.H., Hogue, J.N. & Feminella, J.W. (2000) Develop-
ment and evaluation of predictive models for measuring the biological integrity
of streams. Ecological Applications,
10,14561477.
Hubbell, S.P. & Johnson, L.K. (1977) Competition and nest spacing in a tropical
stingless bee community. Ecology, 58, 949963.
Hunter, M.D., Kozlov, M.V., Itamies, J., Pulliainen, E., Back, J., Kyro, E.M. &
Niemela, P. (2014) Current temporal trends in moth abundance are counter to
predicted effects of climate change in an assemblage of subarctic forest moths.
Global Change Biology, 20, 17231737.
Hurd, P.D. Jr (1979) Superfamily Apoidea. Catalog of Hymenoptera in Amer-
ica north of Mexico.Vol.2 (eds K.V. Krombein, P.D. Hurd Jr, D.R.
Smith & B.D. Burks), pp. 17412209. Smithsonian Institution Press,
Washington, DC.
Inouye, D.W. (1977) Species structure of bumblebee communities in North
America and Europe. The Role of Arthropods in Forest Ecosystems (ed. W.J.
Mattson), pp. 3540. Springer Verlag, New York.
Inouye, D.W. (2008) Effects of climate change on phenology, frost damage, and
floral abundance of montane wildflowers. Ecology, 89,353362.
Jha, S. & Vandermeer, J.H. (2010) Impacts of coffee agroforestry management
on tropical bee communities. Biological Conservation, 143,14231431.
Kearns, C.A. & Oliveras, D.M. (2009) Environmental factors affecting bee diver-
sity in urban and remote grassland plots in Boulder, Colorado. Journal of
Insect Conservation, 13,655665.
Klein, A.M., Vaissiere, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham,
S.A., Kremen, C. & Tscharntke, T. (2007) Importance of pollinators in chang-
ing landscapes for world crops. Proceedings of the Royal Society B-Biological
Sciences, 274, 303313.
Kocher, S.D., Pellissier, L., Veller, C., Purcell, J., Nowak, M.A., Chapuisat, M. &
Pierce, N.E. (2014) Transitions in social complexity along elevational gradients
reveal a combined impact of season length and development time on social
evolution. Proceedings of the Roy al Society B-Biological Sciences, 281,
20140627. doi:10.1098/rspb.2014.0627.
Langenheim, J.H. (1962) Vegetation and environmental patterns in Crested Butte
area, Gunnison County, Colorado. Ecological Monographs, 32,249285.
LeBuhn, G., Griswold, T., Minckley, R., Droege, S., Roulston, T., Cane, J. et al.
(2003) A standardized method for monitoring bee populations The Bee
Inventory (BI) Plot.
Lebuhn, G., Droege, S., Connor, E.F., Gemmill-Herren, B., Potts, S.G., Minck-
ley, R.L. et al. (2013) Detecting Insect Pollinator Declines on Regional and
Global Scales. Conservation Biology, 27,113120.
Lebuhn, G., Droege, S., Connor, E.F., Gemmill-Herren, B., Potts, S.G., Minck-
ley, R.L. et al. (2015) Evidence-based conservation: reply to Tepedino et al.
Conservation Biology, 29,283285.
Lovett, G.M., Burns, D.A., Driscoll, C.T., Jenkins, J.C., Mitchell, M.J., Rustad,
L., Shanley, J.B., Likens, G.E. & Ha euber, R. (2007) Who needs environmen-
tal monitoring? Frontiers in Ecology and the Environment, 5,253260.
Mazerolle, M.J. (2013) AICcmodavg: model selection and multimodel inference
based on (Q)AIC(c). R package version 1.35.
Michener, C.D. (2007) TheBeesoftheWorld, 2nd edn. Johns Hopkins University
Press, Baltimore, MD.
Minckley, R.L., Wcislo, W.T., Yanega, D. & Buchmann, S.L. (1994) Behavior
and phenology of a specialist bee (Dieunomia) and sunflower (Helianthus) pol-
len availability. Ecology, 75, 14061419.
Minteer, B.A., Collins, J.P. & Puschendorf, R. (2014b) Specimen collection: an
essential toolresponse. Science (New York, NY), 344, 816.
Minteer, B.A., Collins, J.P., Love, K.E. & Puschendorf, R. (2014a) Ecology.
Avoiding (re)extinction. Science (New York, NY), 344,260261.
Mitchell, T.B. (1960) Bees of the Eastern United States. North Carolina Agricul-
tural Experiment Station.
Neame, L.A., Griswold, T. & Elle, E. (2013) Pollinator nesting guilds respond dif-
ferently to urban habitat fragmentation in an oak-savannah ecosystem. Insect
Conservation and Diversity, 6,5766.
Novotny,V.,Miller,S.E.,Hulcr,J.,Drew,R.A.I.,Basset,Y.,Janda,M.et al.
(2007) Low beta diversity of herbivorous insects in tropical forests. Nature,
448, 692U698.
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara,
R.B. et al. (2012) vegan: Community Ecology Package.
Olesen, J.M., Bascompte, J., Elberling, H. & Jordano, P. (2008) Temporal
dynamics in a pollination network. Ecology, 89, 15731582.
Ollerton, J., Winfree, R. & Tarrant, S. (2011) How many flowering plants are pol-
linated by animals? Oikos, 120, 321326.
Popic, T.J., Davila, Y.C. & Wardle, G.M. (2013) Evaluation of common methods
for sampling invertebrate pollinator assemblages: net sampling out-perform
pan traps. PLoS ONE, 8, e66665.
Potts, S.G. & Willmer, P. (1997) Abiotic and biotic factors influencing nest-site
selection by Halictus rubicundus, a ground-nesting halictine bee. Ecological
Entomology, 22,319328.
Potts, S.G., Biesmeijer, J.C., Kremen, C., Neumann, P., Schweiger, O. & Kunin,
W.E. (2010) Global pollinator declines: trends, impacts and drivers. Trends in
Ecology & Evolution, 25, 345353.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
10 Z. J. Gezon et al.
R Development Core Team (2011) R: A Language and Environment for Statistical
Computing. R Foundation for Statistical Computing, Vienna, Austria.
Reynoldson, T.B. & Metcalfe-Smith, J.L. (1992) An overview of the assessment
of aquatic ecosystem health using benthic invertebrates. Journal of Aquatic
Ecosystem Health, 1, 295308.
Roulston, T.A.H. & Goodell, K. (2011) The role of resources and risks in regulat-
ing wild bee populations. Annual Review of Entomology, 56,293312.
SAS Institute. (2013) JMP, Version 11.0.0. SAS Institute, Cary, North Carolina,
USA.
Schonberg, L.A., Longino, J.T., Nadkarni, N.M., Yanoviak, S.P. & Gering, J.C.
(2004) Arboreal ant species richness in primary forest, secondary forest, and
pasture habitats of a tropical montane landscape. Biotropica, 36, 402409.
Shepherd, M., Buchmann, S.L., Vaughan, M. & Black, S.H. (2003) Pollinator
Conservation Handb ook: A Guide to Und erstanding, Protecting, and Prov iding
Habitat for Native Pollinator Insects. The Xerces Society, Portland.
Smith, B. & Wilson, J.B. (1996) A consumer’s guide to evenness indices. Oikos,
76,7082.
Statzner, B., Bis, B., Doledec, S. & Usseglio-Polatera, P. (2001) Perspectives for
biomonitoring at large spatial scales: a unified measure for the functional com-
position on invertebrate communities in European running waters. Basic and
Applied Ecology, 2,7385.
Steffan-Dewenter, I., Potts, S.G. & Packer, L. (2005) Pollinator diversity and crop
pollination services are at risk. Trends in Ecology & Evolution, 20,651652.
Steffan-Dewenter, I. & Schiele, S. (2008) Do resources or natural enemies drive
bee population dynamics in fragmented habitats? Ecology, 89, 13751387.
Tepedino, V.J. & Stanton, N.L. (1980) Spatiotemporal variation in phenology
and abundance of oral resources on shortgrass prairie. The Great Basin Natu-
ralist, 40, 197215.
Tepedino, V.J. & Stanton, N.L. (1981) Diversity and competition in bee-plant
communities on short-grass prairie. Oikos, 36,3544.
Tepedino, V.J., Durham, S., Cameron, S.A. & Goodell, K. (2015) Documenting
bee decline or squandering scarce resources. Conservation Biology, 29,280
282.
Wagner, D.H. (1991) The “1-in-20 rule” for plant collectors. Plant Science Bulle-
tin, 37, 11.
Westphal, C., Bommarco, R., Carre, G., Lamborn, E., Morison, N., Petanidou,
T. et al. (2008) Measuring bee diversity in different European habitats and bio-
geographical regions. Ecological Monographs, 78, 653671.
Williams, N.M., Crone, E.E., Roulston, T.H., Minckley, R.L., Packer, L. &
Potts, S.G. (2010) Ecological and life-history traits predict bee species
responses to environmental disturbances. Biological Conservation, 143, 2280
2291.
Wilson, J.S., Griswold, T. & Messinger, O.J. (2008) Sampling bee communities
(Hymenoptera: Apiformes) in a desert landscape: are pan traps sufficient?
Journal of the Kansas Entomological Society, 81,288300.
Wilson, J.S., Messinger, O.J. & Griswold, T. (2009) Variation between bee com-
munities on a sand dune complex in the Great Basin Desert, North America:
implications for sand dune conservation. Journal of Arid Environments, 73,
666671.
Winfree, R. (2010) The conservation and restoration of wild bees. Year in Ecology
and Conservation Biology, 2010, 169197.
Winfree, R., Griswold, T. & Kremen, C. (2007) Effect of human disturbance on
bee communities in a forested ecosystem. Conservation Biology, 21,213223.
Winfree, R., Williams, N.M., Dushoff, J. & Kremen, C. (2014) Species abun-
dance, not diet breadth, drives the persistence of the most linked pollinators as
plant-pollinator networks disassemble. American Naturalist, 183, 600611.
Worm,B.,Hilborn,R.,Baum,J.K.,Branch,T.A.,Collie,J.S.,Costello,C.et al.
(2009) Rebuilding global fisheries. Science, 325, 578585.
Received 22 January 2015; accepted 10 March 2015
Handling Editor: Jana Vamosi
Supporting Information
Additional Supporting Information may be found in the online version
of this article.
Fig. S1. Map of the sampling sites. Treatment refers to sites that were
sampled twice-monthly from 2009 to 2013 (‘repeated’) or only once in
2012 (‘single’).
Fig. S2. Rarefaction curve for the number of morphospecies caught
over the course of the 2012 sampling season.
Fig. S3. Intra-annual variation in bee catch rates (mean bees caught per
hour) for caught-and-released beesvariedfromyeartoyear,withno
consistent trend.
Fig. S4. Patterns of vouchered bee catch rates varied from year to year,
with a trend toward higher catch rates early and late in the season,
except in the year 2012.
Fig. S5. Least Square Means (LSM) from
ANCOVAsofallbees(vou-
chered plus caught-and-relea sed bees) between single-sampled and
repeat-sampled sites.
Fig. S6. Least Square Means (LSM) from nested
ANCOVA testing the
effect of treatment (repeated vs. single sampling) on vouchered bee
catch rate [log(bees/hour + 1)] by sociality.
Fig. S7. Three-dimensional non-metric multidimensional scaling
(NMDS) ordination plots indicating overlap in bee communities when
separated by (a) treatment, (b) river valley, (c) habitat type, and (d)
sample week.
Appendix S1. Sampling site information and locations.
Appendix S2. Species, morphospecies, and functional group designa-
tions used in analyses.
Appendix S3. Table comparing apriorimodels of bee catch rate within
and among years in the repeatedly sampled sites.
Appendix S4. Nested ANCOVAs testing the effect of treatment
(repeated vs. single sampling) on (A) bee catch rate [log(bees/hour)],
(B) rarefied richness, and (C) evenness (measured as E
var
)forallbees
combined (vouchered and catch-and-release).
Appendix S5. Nested
ANCOVA testing the effect of treatment (repeated
vs. single sampling) on vouchered bee catch rate [log(bees/hour + 1)]
by sociality.
© 2015 The Authors. Methods in Ecology and Evolution © 2015 British Ecological Society, Methods in Ecology and Evolution
Effect of lethal sampling on bee populations 11
... This study was conducted in meadows surrounding the Rocky Mountain Biological Laboratory (RMBL) in Gothic, Colorado, USA (elevation 2950 m). The RMBL is located in a sub-alpine region hosting over 150 species of bees and a diverse mixture of perennial flowering plants, many of which rely on bees for successful reproduction [34,35]. This area is particularly vulnerable to climate change as spring temperatures have increased by approximately 1.6°C [36], and spring snowmelt has occurred at a rate of 3.5 days earlier per decade over the last 40 years [35]. ...
... (electronic supplementary material, information S1, [34]). Briefly, during each sampling period, we placed painted bowls of alternating colours (white, fluorescent blue, and florescent yellow and filled with soapy water to attract bees) on the ground every 3 m along two fixed 30 m transects between approximately 09.00 and 17.00 on clear, sunny days. ...
... Briefly, during each sampling period, we placed painted bowls of alternating colours (white, fluorescent blue, and florescent yellow and filled with soapy water to attract bees) on the ground every 3 m along two fixed 30 m transects between approximately 09.00 and 17.00 on clear, sunny days. Bee bowls collect predominately small-to medium-sized bees [34], so we also used a non-lethal method of collecting larger bees (in our case primarily bumblebees) by netting for 2 h at each site (1 h each in the morning and afternoon every day that bowls were out at a site). Bumblebees were captured with a net, transferred to a clear vial, identified based on pile colours using identification guides [40], marked with a non-toxic paint pen to avoid re-capturing, and then released back into the site. ...
Article
Full-text available
Life-history traits, which are physical traits or behaviours that affect growth, survivorship and reproduction, could play an important role in how well organisms respond to environmental change. By looking for trait-based responses within groups, we can gain a mechanistic understanding of why environmental change might favour or penalize certain species over others. We monitored the abundance of at least 154 bee species for 8 consecutive years in a subalpine region of the Rocky Mountains to ask whether bees respond differently to changes in abiotic conditions based on their life-history traits. We found that comb-building cavity nesters and larger bodied bees declined in relative abundance with increasing temperatures, while smaller, soil-nesting bees increased. Further, bees with narrower diet breadths increased in relative abundance with decreased rainfall. Finally, reduced snowpack was associated with reduced relative abundance of bees that overwintered as prepupae whereas bees that overwintered as adults increased in relative abundance, suggesting that overwintering conditions might affect body size, lipid content and overwintering survival. Taken together, our results show how climate change may reshape bee pollinator communities, with bees with certain traits increasing in abundance and others declining, potentially leading to novel plant-pollinator interactions and changes in plant reproduction.
... Furthermore, relative abundances are a community-level pattern and do not give direct measures of species' population-level changes over time (Gotelli et al., 2021). An alternative approach is to conduct standardized sampling, often using passive traps, in the same locations across multiple years (Gezon et al., 2015;Graham et al., 2021;Iserbyt & Rasmont, 2012;Martins et al., 2013;Onuferko et al., 2018). These sampling approaches have the benefit of providing more direct measures of changes in population-level abundances, and potentially for a wide variety of co-occurring species. ...
... However, most standardized sampling does involve collecting a large number of wild bees which can lead to ethical concerns (Gezon et al., 2015;Portman et al., 2020). Also, it can be logistically difficult to continue these types of standardized sampling for a long enough period of time or with enough sample sites to have the statistical power to detect larger population trends (Didham et al., 2020;Lebuhn et al., 2013;Tronstad et al., 2022). ...
Article
Full-text available
Wild bees form diverse communities that pollinate plants in both native and agricultural ecosystems making them both ecologically and economically important. The growing evidence of bee declines has sparked increased interest in monitoring bee community and population dynamics using standardized methods. Here, we studied the dynamics of bee biodiversity within and across years by monitoring wild bees adjacent to four apple orchard locations in Southern Pennsylvania, USA. We collected bees using passive Blue Vane traps continuously from April to October for 6 years (2014–2019) amassing over 26,000 bees representing 144 species. We quantified total abundance, richness, diversity, composition, and phylogenetic structure. There were large seasonal changes in all measures of biodiversity with month explaining an average of 72% of the variation in our models. Changes over time were less dramatic with years explaining an average of 44% of the variation in biodiversity metrics. We found declines in all measures of biodiversity especially in the last 3 years, though additional years of sampling are needed to say if changes over time are part of a larger trend. Analyses of population dynamics over time for the 40 most abundant species indicate that about one third of species showed at least some evidence for declines in abundance. Bee family explained variation in species‐level seasonal patterns but we found no consistent family‐level patterns in declines, though bumble bees and sweat bees were groups that declined the most. Overall, our results show that season‐wide standardized sampling across multiple years can reveal nuanced patterns in bee biodiversity, phenological patterns of bees, and population trends over time of many co‐occurring species. These datasets could be used to quantify the relative effects that different aspects of environmental change have on bee communities and to help identify species of conservation concern. Six years of bee monitoring shows that bee biodiversity is highly dynamic from month to month and across years. Looking across 40 species, 1/3 show evidence of declining abundance over time.
... Pollinator sampling took place on each site once a month, from March to July. Standardized sweep net transects (Westphal et al., 2008) were conducted on the mid-field strips and on the field margins in order to collect bees and hoverflies visiting the flowers, without impacting the local populations (Gezon et al., 2015). The chosen field margin for the sampling was the one with the highest plant diversity around the experimental field. ...
Thesis
The intensification of agriculture with intensive use of inputs, simplification of the landscape, and reduction of semi-natural habitats, is known to contribute to the decline of arthropods. However, arthropods provide essential ecosystem functions and services in agroecosystems and for agricultural activities such as natural regulation of pests and weeds, and pollination. Increasing plant diversity in cultivated fields through the implementation of agroecological practices and infrastructures is a promising approach to promote the presence of arthropods and their associated ecosystem services. The overall objective of the thesis project was to design and experiment an innovative agroecological practice in maize cropping systems, which is easy to insert into existing cropping systems, and which provides an undisturbed habitat for ground-dwelling arthropods and supplementary floral resources for pollinators. The idea of the proposed practice was to capitalize on an already well known practice, the implementation of winter cover crops, by keeping a strip of a winter cover crop in the middle of the field during the whole maize cultivation period. Therefore, two research objectives were defined: i) to measure the impact of the practice on activity-density and diversity of ground-dwelling arthropods, on their dispersion in the cultivated area (spillover), and on the potential pest regulation in the adjacent cultivated area, and ii) to measure the interest of the practice for the conservation of wild pollinators in intensively cultivated landscapes. Field experiments took place on 12 commercial fields of volunteer farmers in 2019 and 2020, in order to take into account the technical and regulatory constraints of farmers in a conventional arable cropping system. Different species groups were surveyed: communities of ground-dwelling arthropods (carabids, spiders, staphylinids and harvestmen), slugs (the main maize pest in the study area), and pollinating insects (bees and hoverflies). Spatio-temporal dynamics of the different natural enemy groups were analysed, predation rates measured with sentinel prey, and the composition of carabid and wild bee communities investigated. Results show that the undestroyed cover crop strips constitute reservoirs of biodiversity, in particular for carabids, spiders, and bees. In the cropped area, no clear effect of the distance from the strip was detected for the different groups of ground-dwelling arthropods, thus no indication of a potential spillover of arthropods into the crop. However, two main carabid species (Poecilus cupreus and Pterostichus melanarius) were more abundant in the vicinity of the strip (10 meters), but not inside the strip, indicating a potential phenomenon of aggregation of these species towards the strip. Predation rates were higher in the strip and seemed to decrease with increasing distance from the strip into the cropped area. Moreover, carabid and bee communities showed to be different from one habitat to another, as well as the distribution of their ecological traits. Thus, the strips can provide complementary habitat and resources for natural enemies and pollinators. Finally, the spontaneous plants of field margins appeared to be essential for oligolectic and less common bees. The results of this thesis show that the conservation of a cover crop strip in the middle of cropped fields can be an effective practice for the conservation of beneficial arthropods in agricultural landscapes, but can also enhance ecosystem services such as pest regulation. Additionally, the results highlight the importance of preserving or even extending perennial semi-natural habitats such as field margins to contribute to biodiversity conservation in arable cropping systems and landscapes.
... populations through the over-trapping of reproductive individuals. In contrast, Gezon et al. (2015) found no evidence of biweekly captures over a 4-year period negatively impacting bee catch rate or the diversity of wild bee communities when 3000 bees were removed per season. However, caution is warranted in interpreting this study as these results are likely an incomplete assessment of the population dynamics of individual taxa and occur within a region with likely many nearby source populations for recolonization. ...
Article
Full-text available
Bees are important pollinators of wild and domesticated flowering plant species. Over the last 30 years, an increasing number of scientific articles have been published on the ecology and conservation of wild bees. To achieve research goals, many studies have pursued the lethal take of wild bees. Although the impact of lethal take for scientific pursuits is likely negligible compared to the negative impacts of human‐mediated phenomena such as climate change, urbanization, and agricultural intensification, it is important to evaluate the history of lethal take on scientific endeavors. In our study, we evaluated a random sample of 30 years of scientific publications on wild bees. Across 1426 surveyed publications, 536 reported the lethal take of wild bees. We found that 61% of these studies lethally captured wild bees primarily for species identification. Furthermore, we determined passive sampling of wild bees resulted in substantially more lethal collections than active methods per study. However, combined approaches of passive and active collection resulted in the greatest lethal take of wild bees per study. Finally, we determined that 64% of the studies did not provide deposition information for their samples, hindering additional research that could be done with them. The increasing availability of video and photographic devices and artificial intelligence approaches to identification, the development of low and noninvasive molecular methods, and the ease of sharing information, allow for a timely discussion on alternative routes and potentially new best practices in bee research. We focus our discussion on alternative methods for minimizing lethal captures for identification purposes and through passive methods, and for maximizing the utility of the data collected. Finally, we provide a framework for continued engagement among researchers and managers to develop strategies that can contribute to reducing our impact on wild bee communities and making the most of collected specimens. To achieve research goals, many studies have pursued the lethal capture of wild bees. We conducted a literature search of last 30 years of scientific research on wild bees and we found that 61% of studies (from a random sample of the total) lethally captured wild bees, primarily for species identification and with passive sampling methods. We then discuss the current available alternative methods for minimizing lethal captures and maximizing the utility of the data collected, and thus, for reducing our impact on wild bee communities.
... The lack of standardized collection methods between historical and current records means it is difficult to separate if the changes in abundance are the result of changes in population size or changes in sampling efforts (Portman et al., 2020). An alternative approach is to conduct standardized sampling, often using passive traps, in the same locations across multiple years (Iserbyt and Rasmont, 2012, Martins et al. 2013, Gezon et al., 2015, Onuferko et al., 2018, Graham et al., 2021. These types of long-term longitudinal studies of bee communities are rare but the existing ones have reported a mix of species that are stable, increasing, and decreasing over time (e.g. ...
Preprint
Full-text available
Wild bees form diverse communities that pollinate plants in both native and agricultural ecosystems making them both ecologically and economically important. The growing evidence of bee declines has sparked increased interest in monitoring bee community and population dynamics using standardized methods. Here, we studied the dynamics of bee biodiversity within and across years by monitoring wild bees adjacent to four apple orchard locations in Southern Pennsylvania, USA. We collected bees using passive Blue Vane traps continuously from April to October for six years amassing over 26,000 bees representing 144 species. We quantified total abundance, richness, diversity, composition, and phylogenetic structure. There were large seasonal changes in all measures of biodiversity with month explaining an average of 72% of the variation in our models. Changes over time were less dramatic with years explaining an average of 44% of the variation in biodiversity metrics. We found declines in all measures of biodiversity especially in the last 3 years. Analyses of population trends over time for the 40 most abundant species indicate that about one third of species showed at least some evidence for declines in abundance. Bee family explained variation in species-level seasonal patterns but we found no consistent family-level patterns in declines, though bumble bees and sweat bees were groups that declined the most. Overall, our results show that season-wide standardized sampling across multiple years can reveal nuanced patterns in bee biodiversity, phenological patterns of bees, and population trends over time of many co-occurring species. These datasets could be used to quantify the relative effects that different aspects of environmental change have on bee communities and to help prioritize conservation efforts.
... Biodiversity knowledge is crucial to propose robust conservation programs in which biodiversity inventories play an important role, especially in understudied well-conserved areas. These studies provide a scientific basis for conservation projects and support science-based decision-making (Lovett et al. 2007;Gezon et al. 2015;Packer and Darla-West 2021;Rosa et al. 2021). Biodiversity surveys allow the compilation of a huge amount of information about species geographic distribution, raw abundance and phenology, and interactions between different organisms -what is crucial to develop conservation strategies (Silveira et al. 2010;Meineke et al. 2018;Rosa et al. 2021). ...
Article
Bee surveys bring valuable information regarding species distribution, phenology, and their interactions with host plants, and thus are important to bee conservation, especially in the present scenario of drastic declines reported worldwide. The Brazilian bee fauna has been surveyed intensively since the late 1960s, but the state-of-the-art of this research topic has not been explored in detail. Our main goal is to analyze the scientific literature through a systematic review of Brazilian bee surveys, describing (i) spatial sampling gaps; (ii) preferred sampling frequency and methods; (iii) species identification and vouchering practices; (iv) gender participation; and (v) how the studies are distributed among scientific journals. Our review includes 213 published studies, of which half targeted the entire bee fauna (Apoidea: Apidae sensu lato) while the other half targeted the orchid bees (Euglossini). Sampling sites are predominant in eastern Brazil, particularly within the Atlantic Forest. Hand netting bees for 1 year long was the most frequent sampling protocol to capture bees. Taxonomic identification and voucher deposition practices were mentioned in most of the papers. Men and women published almost the same number of papers, but women were less frequent as lead or senior authors and appeared less as prolific researchers. Most papers were published in a few Brazilian journals. We highlight that surveys in understudied areas are needed, especially in Amazon. Monitoring bees in previously surveyed locations is also encouraged to investigate temporal changes in bee assemblages, such as population declines, impacts of land-use changes, and effects of climate change.
... However, a disadvantage of passive traps is that they can negatively affect the insect population. Although one study has found that there was not a negative impact on the bee communities (Gezon et al. 2015), another study has found that Russels yellow traps have strong negative effects as these traps can kill bumble bee queens (Kuussaari et al. 2008). These traps rely on the same attraction effect as pan and flight traps, and therefore it seems reasonable to expect a negative effect of these traps under specific circumstances. ...
Article
Full-text available
In order to synthesize changes in pollinating insect communities across space and time, it is necessary to understand whether, and how, sampling methods influence assessments of community patterns. We compared how two common sampling methods-yellow combined flight traps and net sampling-influence our understanding of the species richness, abundance and composition of wild bees and hoverflies, and addressed whether these patterns resulted from potentially biased sampling of individuals or species with different types of functional traits. We sampled bee and hoverfly communities in six sites over three seasons in Saxony-Anhalt, Germany. We captured more species and individuals of bees with traps and more species and individuals of hoverflies with net sampling. However, rarefied richness results were less dramatic between the sampling methods for bees and were not different between the sampling methods for hoverflies. Thus, differences in species richness across sampling methods were mostly due to differences in the number of individuals captured in the different methods. We captured more small-sized bees and hoverflies with traps. We tested if the different methods collected individuals and species with different functional traits, such as nesting preferences, sociality and flower specialization for bees and floral preference, migratory status and habitat preference for hoverflies. For most traits, we collected more individuals but not more species with a certain trait in the different methods. This was mainly due to a high abundance of one species being collected in the different methods. These results suggest that the best methodology depends on the aim of the survey, and that the methods cannot be easily combined into synthesis research. Our results have implications for the development of monitoring schemes for pollinators and for synthesis of trends that can identify threats to pollinators and inform research of pollinator conservation strategies.
... [The effect of yearly sampling of bees is presently unclear. One short-term study of 3 years showed no effect of yearly sampling on diversity of the bee community (Gezon et al., 2015) and another study of 7 years showed no consistent effect on 15 species of xylophilous Hymenoptera (V.J. Tepedino, F. D. Parker, S. Droege, unpublished). Conversely, Gibbs et al. (2017) found a significant decline in two species over 3 years of sampling in orchards. ...
Article
Full-text available
• Continued reports of bee declines have prompted repeated calls for a national monitoring programme in the United States. Here, we argue that such a large-scale surveillance monitoring programme would consume inordinate resources without providing the sought-after data in a meaningful time period. • Surveillance monitoring cannot provide reliable and timely measures of absolute or relative population size, species richness or diversity, host plant and nest site preferences or typical foraging ranges. In addition, the multitude of specimens captured by the passive traps commonly used to surveil overwhelm our identification capacity and, in some cases, may even accelerate species decline. • Conversely, surveillance monitoring should be pursued when data can be collected non-lethally for readily identifiable species or species groups (e.g., bumble bees) or when widely distributed sentinel plants can be observed (e.g., sunflowers). Lethal collections should continue at appropriate temporal intervals but only at a limited number of representative areas with accurate historic records. • We advocate that emphasis be placed on a targeted, natural-history approach aimed at addressing specific hypotheses and exploring solutions to conservation management problems. We already know enough about the drivers of bee decline to formulate and test potential solutions. While additional information is of value in understanding bee diversity, distributions and evolution, we do not need it to begin to take action; we should be focusing on the best ways to help bee species endure rather than attempting to accumulate even more evidence that they are in trouble.
Article
Full-text available
Species identification is currently a strong limitation to wild pollinator studies. It requires killing specimens for laboratory analyses, which can pose ethical issues in some programs of conservation biology and citizen science. The recent development of image-based identification tools using machine learning could challenge the current paradigm of required specimen euthanasia for species identification. However, to be accurate, these new methods call for standardized images or images of precise characters that are difficult or even impossible to obtain on live specimens. To facilitate the acquisition of these pictures, we tested two in-situ CO2 anesthesia protocols using material easily available, even in citizen science programs. We measured the time of anesthesia of 196 flower visitors belonging to the Hymenoptera and Diptera orders. The most efficient protocol enabled us to anesthetize 90 % of the specimens for more than a minute with a marginal mortality (1.5 %). Anesthesia time increased with specimen size in Hymenoptera and decreased with air temperature. Diptera were less sensitive to anesthesia. Further analyses would be required to investigate the potential sublethal effects of these anesthesia. These preliminary results suggest nonetheless that CO2-based anesthesia could help the development of non-lethal methods of wild pollinator identifications.
Chapter
Promoting insect conservation in Australia depends on wider appreciation of insect variety, importance and vulnerability, and that their conservation cannot be pursued comprehensively by professional scientists alone, or by the current financial and other logistic resources available, or that are likely to become available. Continually increased interest and appreciation from the wider populace, including young people, is a key component of rendering insect conservation both socially acceptable and potentially successful. Examples of ‘citizen science’ involvement in insect surveys and related conservation activities have already done much to increase and stimulate conservation in Australia, with attention both to diversity and individual ‘flagship’ species major contributions toward that objective. They are discussed in the context of increasing awareness of insect diversity and understanding the richness and vulnerability of numerous native taxa and their restricted environments, whilst acknowledging that information on insect richness and ecology is still far from complete.
Technical Report
Full-text available
Description Fit linear and generalized linear mixed-effects models. The models and their components are represented using S4 classes and methods. The core computational algorithms are implemented using the 'Eigen' C++ library for numerical linear algebra and 'RcppEigen' ``glue''.
Article
Full-text available
Several diversity-related measures (richness, H′, hierarchical information, species abundance distributions) are used to examine the hypothesis that members of the bee guild of high-altitude, shortgrass prairie compete for limited floral resources. Data on bee and flower populations was collected for two years at two sites. We reasoned that competition should be reflected by: (1) similar species abundances distributions for bees and flowers; (2) strong positive correlations between the total flower abundance of a species and the number of bee visitors; (3) strong positive correlations between bees and flowers in diversity and abundance when samples were taken at frequent intervals. Results are generally inconsistent with the hypothesis that either bees continuously compete for floral resources or vice-versa. Instead, the data support the hypothesis that competition is intermittent due to (1) a spatiotemporally unpredictable resource base, (2) the fact that adult bees are the products of the previous years resource supply rather than that available in the year of emergence. /// Несколько величин, определяющих разнообразне (богатсво, Н́, иерархическая информация, распределение обилия внодов) использованы для проверки гипотезы о том, что пеелы в сообществе высотной низкотравной прерии конкурируют за лимитированные флорнстические ресурсы. Данные по характеристике популяций пчел и цветковых растений собраны в течение 2-х лет в 2-х местообитаниях. Мы предполагаем, что конкуренция зависнт от следующих факторов: 1. сходное распределение обилия определенных вндов у пчел и цветковых; 2. тесная положнтельная межцу общим обилием цветов у одного вида и числом посещений пчел. 3. тесная положительная корреляция величин, характеризующих разнообразие и обилие у пчел и цветов, при взятии проб через короткие интерваы. Пезультаты в целом не сообразуютчя с гнпотезой о наличии у пчел постоянной конкуренции за флорнстические ресурсы, либо ее отсутствии. Вместо этого данные поддерживают гипотезу о наличии прерывнстой конкуренции в пезультате: 1. пространственно-временной нелредсказуемой омене ресурсов и 2. того факта, что наличне взросных пчел отражает скорее величину ресурсов прощых лет, чем их доступность в род выпета имаго.
Article
The density and dispersion patterns of nests of 5 stingless bee species are described for a tract of Costa Rican tropical dry forest. Features of suitable nest sites are analyzed, but it does not appear that nest site availability limits colony density or determines colony dispersion. Rather, food limitation is suggested by a linear relationship between the logarithm of colony biomass and the logarithm of foraging "home range." Colonies were uniformly dispersed intraspecifically in 4 species. The pattern in a 5th species could not be distinguished from a random dispersion. Partial nest counts in 3 more species suggested a clumped dispersion in at least 1 species. The 4 species with uniform nest spacing all use pheromones for recruitment in foraging, whereas the remaining 4 species do not. The 3 most uniformly dispersed species are also intraspecifically aggressive at food sources. We propose that colony spacing is accomplished in these species by aggressive prevention of new colony establishment. The mechanism of colony spacing we propose has as essential elements: pheromone marking of potential nest sites; recruitment of workers; and aggression between workers from rival nests. This mechanism has been observed in operation in at least 1 of the species.