ArticlePDF Available

Abstract and Figures

It is important to understand how biodiversity, including that of rare species, affects ecosystem function. Here we consider this question with regard to pollination. Studies of pollination function have typically focused on pollination of single plant species, or average pollination across plant species, and typically find that pollination depends on a few common species. Here, we used data from 11 plant-bee visitation networks in New Jersey, USA, to ask whether the number of functionally important bee species changes as we consider function separately for each plant species in increasingly diverse plant communities. Using rarefaction analysis, we found the number of important bee species increased with the number of plant species. Overall, 2.5 to 7.6 times more bee species were important at the community scale, relative to the average plant species in the same community. This effect did not asymptote in any of our datasets, suggesting that even greater bee diversity is needed in real world systems. Lastly, on average across plant communities, 25% of bee species that were important at the community scale were also numerically rare within their network, making this study one of the strongest empirical demonstrations to date of the functional importance of rare species.
The number of important pollinator species increases with the number of plant species. (a) Accumulation curves for each of the 11 networks. Points represent the number of pollinator species important to at least one plant species in the full community, and lines represent the accumulation of important pollinator species across levels of plant species richness (i.e., means of rarefied plant communities) where the left end represents the average single plant species, and the right end represents the full plant community. (b) An example of one network’s accumulation curve, now shown together with its null model and 95% CIs. The null model curve represents the expectation if individual pollinators forage randomly across the available plant species, while the observed curve includes biological effects, such as species-specific preferences, morphology, or phenology that led to non-random foraging. (c) Z-scores for each network, representing the strength of the biological effects (complementarity) on the number of pollinator species found to be functionally important in a network, relative to the expectation under random foraging. Z-scores were calculated as the difference between observed and null expectation (large, red vertical bar in b) divided by the standard deviation of the null (small, blue vertical bar in b) at maximum plant richness for each network (i.e., at the endpoints of the curves in a and b). In (a) and (c), the blue line and points (light grey in black and white) represent the experimental garden.
… 
Content may be subject to copyright.
1
Title: Many bee species, including rare species, are important for function of entire plant-1
pollinator networks 2
URL: http://rspb.royalsocietypublishing.org/lookup/doi/10.1098/rspb.2021.2689 3
DOI: http://doi.org/10.1098/rspb.2021.2689 4
Authors: 5
Dylan T. Simpson1 (corresponding author) dylan.simpson@rutgers.com; 425-615-4871 6
Lucia R. Weinman1 lucia.weinman@rutgers.edu 7
Mark A. Genung2,3 mark.genung@louisiana.edu 8
Michael Roswell1,4 mroswell@umd.edu 9
Molly MacLeod1,5 molly.k.macleod@gmail.com 10
Rachael Winfree2 rwinfree@rutgers.edu 11
1 Graduate Program in Ecology & Evolution, 14 College Farm Road, Rutgers University, New 12
Brunswick, NJ 08901, USA 13
2 Department of Ecology, Evolution & Natural Resources, Rutgers University, New Brunswick, 14
New Jersey 08901, USA 15
3 Dept of Biology, University of Louisiana, Lafayette, LA 70503, USA 16
4 Dept. of Entomology, University of Maryland, College Park, MD 20742, USA 17
5 BioMarin Pharmaceutical Inc., Science Communications and Engagement, USA. 18
19
2
Abstract 20
It is important to understand how biodiversity, including that of rare species, affects ecosystem 21
function. Here we consider this question with regard to pollination. Studies of pollination 22
function have typically focused on pollination of single plant species, or average pollination 23
across plant species, and typically find that pollination depends on a few common species. Here, 24
we used data from 11 plant-bee visitation networks in New Jersey, USA, to ask whether the 25
number of functionally important bee species changes as we consider function separately for 26
each plant species in increasingly diverse plant communities. Using rarefaction analysis, we 27
found the number of important bee species increased with the number of plant species. Overall, 28
2.5 to 7.6 times more bee species were important at the community scale, relative to the average 29
plant species in the same community. This effect did not asymptote in any of our datasets, 30
suggesting that even greater bee diversity is needed in real world systems. Lastly, on average 31
across plant communities, 25% of bee species that were important at the community scale were 32
also numerically rare within their network, making this study one of the strongest empirical 33
demonstrations to date of the functional importance of rare species. 34
Keywords: biodiversity; ecosystem function; mutualism; bipartite networks; pollination; rare 35
species 36
37
3
Main Text 38
Introduction 39
Given the rapid loss of global biodiversity [1], it is imperative to understand how decreased 40
biodiversity will affect functioning of natural systems [2]. In particular, ecologists need to 41
understand the role of rare species in ecosystem function, given that rare species are at highest 42
risk of extinction and are the primary focus of conservation [3]. 43
Ecologists’ understanding of biodiversity-ecosystem function (BEF) relationships has evolved as 44
study systems have increasingly resembled natural systems. In experiments, which often focus 45
on single functions within one trophic level, greater biodiversity (specifically, species richness) 46
increases ecosystem function, but function is often maximized at relatively low richness or is 47
driven by high-functioning, dominant species [4–7]. Further work on BEF relationships, 48
however, has highlighted nuance that comes from real-world complexity. In particular, BEF 49
relationships can be mediated by spatiotemporal scale [8–10], the number of functions being 50
considered (i.e., multifunctionality) [11–13], trophic interactions [14–17], and facilitation, 51
including plant-soil feedbacks [18–21]. In these contexts, diversity effects often appear stronger 52
than in simplified experiments (e.g. [8,11]), though this is not always the case [22–25]. 53
An important consequence of studying BEF for a single function, place or time is that these 54
narrow lenses can obscure the functional roles of rare species. Often, common species appear to 55
provide most of the function while rare species appear to contribute relatively little [23,26–29]. 56
This is even true in natural systems; for example, regional-scale analyses show 1% of 57
Amazonian tree species store 50% of the carbon [30], and 2% of bee species provide 80% of 58
crop pollination [31]. However, these examples focus only on carbon storage while omitting 59
4
myriad other ecosystem processes, or only on crop pollination, while omitting the pollination of 60
diverse, wild plant communities. 61
Considering the many dimensions of natural systems has revealed some ways that rare species 62
can be important to ecosystem functioning. For example, rare species can contribute 63
disproportionately to functional trait diversity [32–34], which may facilitate multifunctionality 64
[35,36] or maintain function across variable environments [37–39]. Additionally, rare species 65
can contribute disproportionately to invasion resistance and food web stability [40–43]. All of 66
these examples suggest that the (observed) importance of rare species can depend on which 67
function(s) are measured and at which scale(s). Thus, it is important that BEF relationships be 68
studied in their most relevant, real-world contexts. 69
Animal-mediated pollination is used by ~88% of plant species [44] and is one of the model 70
systems for BEF research [2], but the study of plant communities in this context has been limited 71
[45]. Most studies quantifying function within plant-pollinator networks have been simplified to 72
either focus narrowly on the pollination of a single plant species (typically of a crop 73
monoculture, e.g. [31,46,47]), or focus coarsely on average pollination across plant species (e.g. 74
[48]). In nature, however, even in one time and place, pollination must be provided to many 75
species simultaneously. Similarly, plant-pollinator interaction networks have been well-studied 76
in ecological contexts (e.g., with respect to community stability or species interactions [49–52]), 77
but network-level data have rarely been used in a BEF context. The relationship between 78
network structure and ecosystem function has been explored theoretically [45,53], but empirical 79
studies are rare (see [6,54]), especially in natural communities. Thus, despite the attention paid to 80
plant-pollinator networks broadly, the simple question of how many pollinator species are 81
needed to pollinate natural plant communities has yet to be addressed. 82
5
The number of pollinator species needed to pollinate a plant community will depend on the 83
extent of differences among pollinator species in the plant species they visit (i.e., functional 84
redundancy versus complementarity) (Figure 1). On the one hand, most plant-pollinator 85
networks exhibit some degree of nestedness, such that rare or specialist pollinators tend to 86
interact with abundant, generalist plant species (and vice versa) [55,56]. The more nested a 87
network is, the more redundant pollinator species will tend to be, because a few abundant 88
generalists will dominate pollination across plant species (Figure 1b,d). On the other hand, 89
networks are not perfectly nested and some degree of functional complementarity among 90
pollinators is also common [57]. The more complementary pollinator species are in their plant 91
use, the greater need there will be for pollinator diversity at the scale of the entire network 92
(Figure 1c,d) [53]. 93
Complementarity within plant-pollinator networks can also create a functional role for pollinator 94
species that are rare within their network. Pollination of a single plant species, at least in one 95
time and place, tends to be dominated by a few abundant pollinator species, while locally rare 96
species contribute relatively little [58,59] (but see [60]). The situation might be very different, 97
however, when function is considered for each species in a plant community. For example, a 98
pollinator species that is rare within the community could still be an important pollinator of a 99
particular plant species if it is among the most frequent visitors to that plant (Figure 1c). In this 100
situation, locally rare pollinator species could still be important for pollination of the entire plant 101
community, but this effect would be missed in studies in which function is measured for only a 102
single plant species or is averaged across plant species. 103
Here, we use data from 11 plant-bee networks to ask how many pollinator species are needed to 104
pollinate all of the plants in each network. To control for sampling effects and distinguish effects 105
6
of stochasticity from those of complementarity among bee species, we used a randomization-106
based null model. Specifically, we ask 1) What is the relationship between the number of plant 107
species in a network and the number of bee species important for pollinating them? And 2) How 108
important are rare bee species to pollination? 109
Methods 110
Network data 111
We used 11 quantitative plant-bee network datasets collected by our lab in New Jersey, USA 112
[61–63] (Text S1; Figure S1; Table S1). Each dataset quantifies bee visitation to each species of 113
a plant community, as observed in a single site in a single year. We chose datasets collected in 114
one site and year so that differences in plant use by bees could not be driven by spatial or annual 115
turnover in the bee community. Ten of these datasets were collected in natural or semi-natural 116
meadows, while one was a planted field experiment in which each plant species was maintained 117
at equal abundance. 118
Most of these networks include plant species on which few individual bees were observed. To 119
limit our analyses to plant species for which we could be relatively confident of the visiting bee 120
community, we excluded plant species with fewer than 20 observed plant-bee interactions 121
(Tables S1-S3). This meant excluding a mean of 54% of plant species (range = 0 – 83% across 122
networks), but only 8.0% of bee species (range = 0 – 33%) and 9.5% of individual plant-bee 123
interactions (range = 0 – 27%). For the 11 datasets as analyzed, plant species richness varied 124
from 6 to 23, bee species richness varied from 22 to 86, and total individual plant-bee 125
interactions varied from 227 to 4513. In total, the analyzed datasets included 70 plant species and 126
7
173 bee species, with 20943 total observed interactions, and 1479 unique species-species pairs 127
across networks. 128
Analysis 129
To start, we identified the most functionally important bee species for each plant species within 130
each network. We used interaction frequency (i.e., the number of individual bees of a species 131
that were collected from a given plant species) as a proxy for function, and defined 'functionally 132
important' bee species as those that contributed a threshold percent of visits to at least one plant 133
species in their network [31]. We focus on results based on using a 5% threshold (as used by 134
Kleijn et al. 2015) but, to test the sensitivity of our results to our choice in threshold, we repeat 135
the analyses across thresholds from 2.5% to 10% (see Texts S2-S3 and Figures S2-S3 for further 136
discussion). Although visitation frequency can be an incomplete proxy for pollination function, 137
we believe it is adequate in this case. Technically, a pollinator’s contribution to function also 138
depends also on its effectiveness (per-visit pollen deposition) and efficiency (essentially the 139
‘quality’ of pollination), and there are examples of frequent insect visitors being poor pollinators 140
[64,65]. However, plants’ most frequent floral visitors are typically their most important 141
pollinators, and this relationship is especially pronounced for bees, which are the focus of this 142
study [31,64,66]. Also, on a practical level, it would not have been possible to measure per-visit 143
function for the 1479 unique plant-pollinator interactions in our data sets. Lastly, although plants 144
and bees are mutually dependent, we chose to focus on the role of bee diversity in providing 145
pollination, rather than the role of plant diversity in supporting bees. This was in part to expand 146
on existing biodiversity-ecosystem function literature [2], and also because it is more appropriate 147
given our data, which represent the bees that visit a given plant community. 148
8
149
What is the relationship between the number of plant species in a network and the number of bee 150
species important for pollinating them? 151
Within networks, we performed rarefaction to relate the number of important bee species to plant 152
species richness. More specifically, we subset the observations in each network to generate plant 153
communities of varying richness and counted the number of bee species important to at least one 154
plant species in that set. Thus, just as site-based rarefaction measures the accumulation of new 155
species with additional sites, we measured the accumulation of important bee species with 156
additional plant species. We included every possible level of richness for the network (i.e., from 157
1 to n species) and up to 1000 unique (and random) combinations of plant species per richness 158
level. In instances in which there were ≤ 1000 combinations of plant species, we included all 159
possible combinations. We then took the mean number of important bees across combinations of 160
plant identity for each level of plant richness. 161
We represent results from this analysis as accumulation curves in which the mean number of bee 162
species important to at least one plant species is plotted against the number of plant species in the 163
community (sensu [8,9,67,68]; Figure 1d). Greater complementarity among bee species should 164
result in lower values for single plant species (indicating higher specialization by bees at the 165
plant species level) and/or steeper slopes (indicating greater turnover of important bee species 166
among plants). 167
The slopes observed in these curves will also be due, at least in part, to stochasticity. That is, 168
even if there were no biological differences among bee species in terms of the plants they visit, 169
they will visit plant species at different frequencies due to chance (i.e. sampling error) [51]. 170
9
Similarly, we will observe differences in visitation rates due to human sampling error. As a 171
result, any observed complementarity effect should be a combination of biology and 172
stochasticity. To account for these stochastic effects, we created a randomization-based null 173
model to define an expectation under a scenario of no biological complementarity. This null 174
model assumes that there are no underlying differences among bee species, but rather that 175
individual bees forage randomly across all the plant species in their network. 176
To generate the null expectation, we maintained the total number of observations of each plant 177
species, but assigned interactions by random draw (with replacement) from the network-wide 178
bee-species abundance distribution. Said another way, the model maintained the empirical 179
number of bee visits to each plant species (row sums of the plant-bee matrix), but resampled 180
individual bee interactions with probabilities proportional to each bee species’ relative 181
abundance (column sums). We generated 999 null datasets per network [69] and then, for each of 182
these datasets, we estimated the mean number of important bee species for each level of plant 183
species richness. 184
In the Results, we report three metrics for each network. First, we calculate the change in the 185
number of important bee species recorded for the average single plant species versus for the 186
entire network (i.e., all plant species). This metric shows how the need for bee diversity increases 187
with the number of plant species considered, when both the stochastic and the biological 188
components of that increase are included. Second, we compare the observed number of 189
important bee species to the inner 95th percentile of what was predicted by the null model. 190
Observed values beyond the inner 95th percentile were considered significantly different than 191
what would be expected under random foraging, suggesting that biological effects increase the 192
functional complementarity among bee species and contribute to the need for biodiversity. Third, 193
10
we calculate a standardized effect size (a Z-score) for each network that represents the magnitude 194
of any non-stochastic (i.e., biological) effects on the number of important bee species in that 195
network. Z-scores were calculated as the difference between the observed value and the null 196
prediction, divided by the standard deviation of the null (i.e. (observed – null)/sdnull), where all 197
three values are calculated at maximum plant species richness (i.e., using all the plant species in 198
the network). Thus, the Z-scores measure the strength of biological effects, such as niche 199
partitioning and bee specialization, in driving the need for bee diversity, and express this effect 200
in units of standard deviations of the null distribution. 201
Lastly, because our 11 networks varied in the number of plant species they contained, we also 202
examined the role of plant species richness across (rather than within) networks. Specifically, we 203
looked at Pearson’s correlation between the number of plant species in a network and each of the 204
measures above, as well as simply the total number of bee species that were important to at least 205
one plant species in that network. 206
How important are rare bee species to pollination? 207
Here, we measured how many important bee species (i.e., functionally important to at least one 208
plant species) in each network were also rare within that network. While rarity can be defined in 209
many ways (e.g. [70,71]), we focus simply on local rarity – i.e. species with low relative 210
abundance – which is how rarity has been typically considered in the BEF literature (e.g. 211
[36,42]). This means we do not treat rarity as an intrinsic trait of a species; by our definition, a 212
species could be rare in one community and common in another. In the main text, we focus on an 213
analysis in which rarity was defined as any bee species representing < 1% of all bee observations 214
in its network (sensu [72,73]). However, because any definition of rarity is arbitrary, we also 215
11
repeat the analysis across rarity thresholds of 0.5% to 1.5%. Finally, because rare species may 216
occasionally appear important just due to sampling effects, we use our null model to compare our 217
observed results to the null expectation under random foraging. 218
All our analyses were performed in R (3.6.3), using packages parallel (3.6.3) and pbapply (1.4-219
0). Data management was done with tidyverse (1.3.1) and lubridate (1.7.10). All data and stand-220
alone code needed to re-create our analysis are available in the supplement. 221
Results 222
What is the relationship between the number of plant species in a network and the number of 223
bee species important for pollinating them? 224
Within networks, the number of functionally important bee species increased rapidly with plant 225
species richness (Figures 2, S4). Comparing the average single plant species with their 226
respective communities (i.e., comparing the starting and ending points of the accumulation 227
curves), the number of important bee species increased 2.5 to 7.6-fold (Figure 2a). The 228
accumulation curves of functionally important species were mostly non-saturating and rose 229
beyond the inner 95th percentile of the null in all but two of the smallest networks (Figures 2a-b, 230
S5). Z-scores, which measure the effect of complementarity on the number of important bee 231
species relative to the expectation under random foraging, ranged from 1.1 to 12.2 (Figure 2c). 232
Across networks, these results were each associated with plant species richness (Figure 2). There 233
was a strong correlation between the total number of plant species in a network and i) the 234
observed number of important bee species in that network (r = 0.92, p < 0.001), ii) the factor 235
increases in the number of important bee species, relative to a single plant species (r = 0.95, p < 236
12
0.001), and iii) the Z-score (r = 0.95, p < 0.001). These results were also robust to our choice of 237
threshold for defining functional importance. While the absolute number of important bee 238
species decreased under a higher, less inclusive threshold, the factor-differences between single 239
plant species and their respective communities, and the associated Z-scores (i.e., the relative 240
effect sizes), actually increased under higher thresholds (Text S3; Figure S5, S6). 241
How important are rare bee species to pollination? 242
Of the bee species that are functionally important to at least one plant species in a given network, 243
a mean of 25% (range = 0-52%) were rare within that network (Figures 3, S7, S8). More rare 244
bees were important in more plant-rich networks (r = 0.97, p < 0.001), and this number was 245
significantly greater than the null expectation in all but the smallest networks. As would have to 246
be the case, the proportion of important bee species that are rare decreases with more 247
conservative thresholds (higher thresholds for importance, and lower thresholds for rarity) 248
(Figures S9, S10). Even with the most conservative combination of thresholds, though, a mean of 249
7.8% and as many as 25.0% of the important bee species were rare within their network. 250
Discussion 251
By focusing on the pollination of individual plant species rather than plant communities, 252
ecologists have likely underestimated the importance of pollinator diversity to pollination 253
function in nature. Here, we show that the number of functionally important bee species 254
increases rapidly as we expand from considering one to many plant species (Figure 2). Up to 255
seven times more bee species made important contributions at the community scale, as compared 256
with any single plant species. The number of important bee species increased with the number of 257
plant species in a community and did not asymptote in any of our datasets (Figure 2), suggesting 258
13
that even more bees are important in nature. This increased role of biodiversity results from 259
complementary floral use among bee species. Complementarity in resource use among species is 260
a well-known mechanism through which biodiversity increases ecosystem function in 261
experiments [74], yet the primary way species partition resources within a mutualist network 262
partitioning the partners with which they interact – is invisible when function is measured for 263
only a single partner species or averaged across species. Thus, it has largely been overlooked up 264
to now. 265
Perhaps our most striking finding is that, when the whole plant community was considered, rare 266
bee species were frequently important to function. Regardless of the exact thresholds we used to 267
define importance and rarity, a substantial portion of the functionally important bee species in 268
our analyses were also rare within their community (means of 8-45%, across thresholds; Figures 269
3, S10). This result extends previous work that has suggested rare species could be important, but 270
measured their functional role less directly. For example, rare species have been valued because 271
they contribute to functional trait diversity [32,33,75] and because they could become abundant, 272
and thus functionally important, at other places or times (i.e. insurance effects) [10,37,38]. In 273
contrast, we demonstrated a direct and immediate contribution of locally rare species. This 274
contribution did not depend on rare species making disproportionately large contributions to 275
function (i.e. keystone effects [76,77]), which is another commonly cited way for rare species to 276
be important. Instead, we found rare bees to be important because they filled distinct functional 277
roles [35,77], in this case pollinating different plant species. Mechanistically, our approach of 278
considering the pollination of many plant species is akin to measuring multiple ecosystem 279
functions (i.e. ‘multifunctionality’), where it has likewise been found that locally rare species can 280
provide functions not provided by other, more common species [35]. Both types of findings 281
14
suggest that many more species are important to ecosystem function in complex natural 282
communities, where the number of plant species and ecosystem functions greatly exceed what 283
can be measured by researchers. 284
Floral specialization by pollinators is a well-described phenomenon [57,78], and so it may seem 285
obvious that more plant species would require more pollinator species. Findings from network 286
ecology, however, might predict the opposite. In particular, because plant-pollinator networks 287
are typically nested [55,56], one might expect that abundant generalist pollinators would be 288
responsible for most of the pollination across plant species, with rare or specialist species being 289
largely redundant (Figure 1b,d). Indeed, abundant bees in our study did provide more flower 290
visits than rare bees (by definition), and so did have higher average contributions and were 291
important to more plant species (Figure S8). Yet, if we were to only consider bee species’ 292
average contributions across plant species, we would be ignoring the needs of those plant species 293
that were visited primarily by less abundant bees (Figure S8). Our contribution in this paper is to 294
consider the pollinators needed by the whole plant community, rather than just single plant 295
species or the average plant species, and thereby to reveal the important role played by bee 296
species that are rare at the community scale. Of course, this assumes the function of pollination is 297
simply to support the plant community, and that plant species are equally important. If the 298
greater function of interest is, for example, plant biomass, then these interactions between rarer 299
plant and bee species may be of less consequence. 300
The extent to which our results extend to mutualist systems other than pollination networks 301
likely depends on the extent of complementarity in those systems. Pollination networks tend to 302
be relatively specialized (i.e., high complementarity). Other systems with similarly high levels of 303
specialization (e.g., ant-myrmecophyte networks) might behave similarly, while the importance 304
15
of partner-species diversity may be lower in systems with relatively low specialization (e.g. seed-305
dispersal networks) [57]. There is also already evidence that interaction complementarity in 306
plant-mycorrhizal networks lends an effect of fungal diversity on plant growth [20]. Thus, our 307
study is neither the first nor final word, but is further evidence that we should consider 308
biodiversity-ecosystem function relationships in the context of real-world interaction networks. 309
Because our study was observational, we cannot know what would happen if particular bee 310
species were lost from our networks. In particular, we do not know the pollen limitation status of 311
the plants in our networks, which means we cannot predict how their reproduction would be 312
impacted by some level of pollinator loss. Nor can we predict how the network might restructure 313
after species loss. On the one hand, even a plant that is not currently pollen limited could become 314
so following loss of a dominant pollinator. On the other hand, pollinator species’ preferences are 315
often dynamic [79,80], which should lend resilience to species loss [81]. That is, following the 316
loss of a plant’s dominant pollinator species, other pollinators might shift or expand their diets, 317
which could compensate for the loss [82]. However, increased pollinator generalization 318
following the loss of a competitor can also decrease pollination quality due to increased 319
interspecific pollen transfer [79,83,84]. Future research should work to determine which of these 320
processes are dominant in determining pollination function (i.e., plant reproductive success) in 321
the face of species loss. More broadly, understanding function within mutualistic networks will 322
require understanding the extent to which interactions are fixed or plastic, and whether changes 323
to network structure following species loss affect function for the remaining species. 324
Altogether, our results highlight the many dimensions of ecosystem function, and the importance 325
of considering real-world complexity for understanding biodiversity-ecosystem function 326
relationships in nature. In particular, mechanisms governing BEF relationships in nature may be 327
16
invisible in small-scale or simplified study systems [16,85]. As a result, studying function at too 328
small a scale or in too simple a system may lead us to underestimate the number of species 329
needed for function in nature. For instance, despite positive biodiversity effects [46,74,86], 330
function at local scales often relies on relatively few species because of dominance [4,7,23,29]. 331
Yet, because of species turnover, far more species are needed to maintain function across broader 332
spatiotemporal scales [8,10,37,87,88]. Similarly, more species are needed to maintain multiple 333
functions simultaneously than for any function alone because of functional complementarity 334
[9,13,89,90]. Here, we demonstrate an analogous role of biodiversity in mutualist networks: even 335
for a single function in a single place and time, many more species are needed to maintain 336
function across a network than for any one partner species alone. Real-world ecosystems depend 337
on many functions operating across broad spatiotemporal scales [85] and, like pollination, many 338
of these functions are realized through mutualist interactions [91]. In light of this, our results 339
suggest that biodiversity may be even more important for real-world function than previously 340
supposed. 341
342
17
Data and code: The analysis in this paper uses 11 plant-bee network datasets collected by our 343
lab group. For convenience, these data, along with code to reproduce our analysis and figures, 344
are included here as electronic supplementary material. Any use of these data, however, should 345
cite the original papers. 346
Author contributions: DTS and RW conceived of study in collaboration with LW and MG. 347
DTS developed and performed analyses. DTS drafted the manuscript with input from RW, and 348
all authors contributed to revision. MR and MM contributed data. 349
Acknowledgements: We kindly thank Charlie Nicholson for discussion and review of an earlier 350
version of this manuscript. 351
References 352
1. Rockström J et al. 2009 A safe operating environment for humanity. Nature 461|24, 472–353
475. 354
2. Cardinale BJ et al. 2012 Biodiversity loss and its impact on humanity. Nature 486, 59–67. 355
(doi:10.1038/nature11148) 356
3. Gaston K. 1994 Rarity. Springer Netherlands. 357
4. Cardinale BJ, Srivastava DS, Duffy JE, Wright JP, Downing AL, Sankaran M, Jouseau C. 358
2006 Effects of biodiversity on the functioning of trophic groups and ecosystems. Nature 359
443, 989–992. (doi:10.1038/nature05202) 360
5. Jonsson M, Malmqvist B. 2003 Mechanisms behind positive diversity effects on 361
ecosystem functioning: Testing the facilitation and interference hypotheses. Oecologia 362
134, 554–559. (doi:10.1007/s00442-002-1148-5) 363
18
6. Fründ J, Dormann CF, Holzschuh A, Tscharntke T. 2013 Bee diversity effects on 364
pollination depend on functional complementarity and niche shifts. Ecology 94, 2042–365
2054. (doi:10.1890/12-1620.1) 366
7. Smith MD, Knapp AK. 2003 Dominant species maintain ecosystem function with non-367
random species loss. Ecol. Lett. 6, 509–517. 368
8. Winfree R, Reilly JR, Bartomeus I, Cariveau DP, Williams NM, Gibbs J. 2018 Species 369
turnover promotes the importance of bee diversity for crop pollination at regional scales. 370
Science (80-. ). 359, 791–793. (doi:10.1126/science.aao2117) 371
9. Isbell F et al. 2011 High plant diversity is needed to maintain ecosystem services. Nature 372
477, 199–202. (doi:10.1038/nature10282) 373
10. Allan E, Weisser W, Weigelt A, Roscher C, Fischer M, Hillebrand H. 2011 More diverse 374
plant communities have higher functioning over time due to turnover in complementary 375
dominant species. Proc. Natl. Acad. Sci. 108, 17034–17039. 376
(doi:10.1073/pnas.1104015108) 377
11. Zavaleta ES, Pasari JR, Hulvey KB, Tilman GD. 2010 Sustaining multiple ecosystem 378
functions in grassland communities requires higher biodiversity. Proc. Natl. Acad. Sci. U. 379
S. A. 107, 1443–1446. (doi:10.1073/pnas.0906829107) 380
12. Gamfeldt L, Hillebrand H, Jonsson PR. 2008 Multiple functions increase the importance 381
of biodiversity for overall ecosystem functioning. Ecology 89, 1223–1231. 382
13. Fanin N, Gundale MJ, Farrell M, Ciobanu M, Baldock JA, Nilsson MC, Kardol P, Wardle 383
DA. 2018 Consistent effects of biodiversity loss on multifunctionality across contrasting 384
19
ecosystems. Nat. Ecol. Evol. 2, 269–278. (doi:10.1038/s41559-017-0415-0) 385
14. Peralta G, Frost CM, Rand TA, Didham RK, Tylianakis JM. 2014 Complementarity and 386
redundancy of interactions enhance attack rates and spatial stability in host-parasitoid food 387
webs. Ecology 95, 1888–1896. (doi:10.1890/13-1569.1) 388
15. Poisot T, Mouquet N, Gravel D. 2013 Trophic complementarity drives the biodiversity-389
ecosystem functioning relationship in food webs. Ecol. Lett. 16, 853–861. 390
(doi:10.1111/ele.12118) 391
16. Eisenhauer N et al. 2019 A multitrophic perspective on biodiversity – ecosystem 392
functioning research. Adv. Ecol. Res. 61, 1–54. (doi:10.1016/bs.aecr.2019.06.001) 393
17. Hines J et al. 2015 Towards an integration of biodiversity-ecosystem functioning and food 394
web theory to evaluate relationships between multiple ecosystem services. 1st edn. 395
Elsevier Ltd. (doi:10.1016/bs.aecr.2015.09.001) 396
18. Wright AJ, Wardle DA, Callaway R, Gaxiola A. 2017 The Overlooked Role of 397
Facilitation in Biodiversity Experiments. Trends Ecol. Evol. 32, 383–390. 398
(doi:10.1016/j.tree.2017.02.011) 399
19. Van Der Heijden MGA. 2010 Mycorrhizal fungi reduce nutrient loss from model 400
grassland ecosystems. Ecology 91, 1163–1171. (doi:10.1890/09-0336.1) 401
20. Van Der Heijden MGA, Klironomos JN, Ursic M, Moutoglis P, Streitwolf-Engel R, 402
Boller T, Wiemken A, Sanders IR. 1998 Mycorrhizal fungal diversity determines plant 403
biodiversity, ecosystem variability and productivity. Nature 396, 69–72. 404
(doi:10.1038/23932) 405
20
21. Eisenhauer N, Reich PB, Scheu S. 2012 Increasing plant diversity effects on productivity 406
with time due to delayed soil biota effects on plants. Basic Appl. Ecol. 13, 571–578. 407
(doi:10.1016/j.baae.2012.09.002) 408
22. Chisholm RA et al. 2013 Scale-dependent relationships between tree species richness and 409
ecosystem function in forests. J. Ecol. 101, 1214–1224. (doi:10.1111/1365-2745.12132) 410
23. Winfree R, Fox JW, Williams NM, Reilly JR, Cariveau DP. 2015 Abundance of common 411
species, not species richness, drives delivery of a real-world ecosystem service. Ecol. Lett. 412
18, 626–635. (doi:10.1111/ele.12424) 413
24. Gamfeldt L, Roger F. 2017 Revisiting the biodiversity-ecosystem multifunctionality 414
relationship. Nat. Ecol. Evol. 1. (doi:10.1038/s41559-017-0168) 415
25. van der Plas F. 2019 Biodiversity and ecosystem functioning in naturally assembled 416
communities. Biol. Rev. (doi:10.1111/brv.12499) 417
26. Maas B, Tscharntke T, Saleh S, Dwi Putra D, Clough Y. 2015 Avian species identity 418
drives predation success in tropical cacao agroforestry. J. Appl. Ecol. 52, 735–743. 419
(doi:10.1111/1365-2664.12409) 420
27. Lohbeck M, Bongers F, Martinez-Ramos M, Poorter L. 2016 The importance of 421
biodiversity and dominance for multiple ecosystem functions in a human-modifed tropical 422
landscape. Ecology 97, 2772–2779. (doi:10.1002/ecy.1499) 423
28. Gaston KJ. 2011 Common Ecology. Bioscience 61, 354–362. 424
(doi:10.1525/bio.2011.61.5.4) 425
29. Gaston KJ, Cox DTC, Canavelli SB, García D, Hughes B, Maas B, Martínez D, Ogada D, 426
21
Inger R. 2018 Population Abundance and Ecosystem Service Provision: The Case of 427
Birds. Bioscience 68, 264–272. (doi:10.1093/biosci/biy005) 428
30. Fauset S et al. 2015 Hyperdominance in Amazonian forest carbon cycling. Nat. Commun. 429
6, 1–9. (doi:10.1038/ncomms7857) 430
31. Kleijn D et al. 2015 Delivery of crop pollination services is an insufficient argument for 431
wild pollinator conservation. Nat. Commun. 6, 7414. (doi:10.1038/ncomms8414) 432
32. Mouillot D et al. 2013 Rare species support vulnerable functions in high-diversity 433
ecosystems. PLoS Biol. 11. (doi:10.1371/journal.pbio.1001569) 434
33. Jain M et al. 2014 The importance of rare species: A trait-based assessment of rare species 435
contributions to functional diversity and possible ecosystem function in tall-grass prairies. 436
Ecol. Evol. 4, 104–112. (doi:10.1002/ece3.915) 437
34. Violle C, Thuiller W, Mouquet N, Munoz F, Kraft NJB, Cadotte MW, Livingstone SW, 438
Mouillot D. 2017 Functional Rarity: The Ecology of Outliers. Trends Ecol. Evol. 32, 356–439
367. (doi:10.1016/j.tree.2017.02.002) 440
35. Jousset A et al. 2017 Where less may be more: How the rare biosphere pulls ecosystems 441
strings. ISME J. 11, 853–862. (doi:10.1038/ismej.2016.174) 442
36. Soliveres S et al. 2016 Locally rare species influence grassland ecosystem 443
multifunctionality. Philos. Trans. R. Soc. B Biol. Sci. 371. (doi:10.1098/rstb.2015.0269) 444
37. Loreau M, Mouquet N, Gonzalez A. 2003 Biodiversity as spatial insurance in 445
heterogeneous landscapes. Proc. Natl. Acad. Sci. U. S. A. 100, 12765–12770. 446
(doi:10.1073/pnas.2235465100) 447
22
38. Reich PB, Tilman D, Isbell F, Mueller K, Hobbie SE, Flynn DFB, Eisenhauer N. 2012 448
Impacts of biodiversity loss escalate through time as redundancy fades. Science (80-. ). 449
336, 589–592. (doi:10.1126/science.1217909) 450
39. Gerhard M, Schlenker A, Hillebrand H, Striebel M. 2021 Environmental stoichiometry 451
mediates phytoplankton diversity effects on communities’ resource use efficiency and 452
biomass. J. Ecol. , 1–13. (doi:10.1111/1365-2745.13811) 453
40. Zavaleta ES, Hulvey KB. 2004 Realistic species losses disproportionately reduce 454
grassland resistance to biological invaders. Science (80-. ). 306, 1175–1177. 455
(doi:10.1126/science.1102643) 456
41. Lyons KG, Schwartz MW. 2001 Rare species loss alters ecosystem function - Invasion 457
resistance. Ecol. Lett. 4, 358–365. (doi:10.1046/j.1461-0248.2001.00235.x) 458
42. Bracken MES, Low NHN. 2012 Realistic losses of rare species disproportionately impact 459
higher trophic levels. Ecol. Lett. 15, 461–467. (doi:10.1111/j.1461-0248.2012.01758.x) 460
43. Säterberg T, Jonsson T, Yearsley J, Berg S, Ebenman B. 2019 A potential role for rare 461
species in ecosystem dynamics. Sci. Rep. 9, 11107. (doi:10.1038/s41598-019-47541-6) 462
44. Ollerton J, Winfree R, Tarrant S. 2011 How many flowering plants are pollinated by 463
animals? Oikos 120, 321–326. (doi:10.1111/j.1600-0706.2010.18644.x) 464
45. Schleuning M, Fründ J, García D. 2015 Predicting ecosystem functions from biodiversity 465
and mutualistic networks: An extension of trait-based concepts to plant-animal 466
interactions. Ecography (Cop.). 38, 380–392. (doi:10.1111/ecog.00983) 467
46. Hoehn P, Tscharntke T, Tylianakis JM, Steffan-Dewenter I. 2008 Functional group 468
23
diversity of bee pollinators increases crop yield. Proc. R. Soc. B Biol. Sci. 275, 2283–469
2291. (doi:10.1098/rspb.2008.0405) 470
47. Garibaldi LA et al. 2013 Wild Pollinators Enhance Fruit Set of Crops Regardless of 471
Honey Bee Abundance. Science (80-. ). 339, 1608–1611. (doi:10.1126/science.1230200) 472
48. Kaiser-Bunbury CN, Muff S, Memmott J, Müller CB, Caflisch A. 2010 The robustness of 473
pollination networks to the loss of species and interactions: A quantitative approach 474
incorporating pollinator behaviour. Ecol. Lett. 13, 442–452. (doi:10.1111/j.1461-475
0248.2009.01437.x) 476
49. Bascompte J, Jordano P, Olesen JM. 2006 Asymmetric coevolutionary networks facilitate 477
biodiversity maintenance. Science (80-. ). 1, 431–433. (doi:10.1126/science.1123412) 478
50. Bastolla U, Fortuna MA, Pascual-García A, Ferrera A, Luque B, Bascompte J. 2009 The 479
architecture of mutualistic networks minimzies competition and increases biodiversity. 480
Nature 458, 1018–1020. (doi:10.1038/nature07950) 481
51. Vázquez DP, Blüthgen N, Cagnolo L, Chacoff NP. 2009 Uniting pattern and process in 482
plant-animal mutualistic networks: A review. Ann. Bot. 103, 1445–1457. 483
(doi:10.1093/aob/mcp057) 484
52. Memmott J, Waser NM, Price M V. 2004 Tolerance of pollination networks to species 485
extinctions. Proc. R. Soc. B Biol. Sci. 271, 2605–2611. (doi:10.1098/rspb.2004.2909) 486
53. Blüthgen N, Klein AM. 2011 Functional complementarity and specialisation: The role of 487
biodiversity in plant-pollinator interactions. Basic Appl. Ecol. 12, 282–291. 488
(doi:10.1016/j.baae.2010.11.001) 489
24
54. Kaiser-Bunbury CN, Mougal J, Whittington AE, Valentin T, Gabriel R, Olesen JM, 490
Blüthgen N. 2017 Ecosystem restoration strengthens pollination network resilience and 491
function. Nature 542, 223–227. (doi:10.1038/nature21071) 492
55. Bascompte J, Jordano P, Melián CJ, Olesen JM. 2003 The nested assembly of plant-493
animal mutualistic networks. Proc. Natl. Acad. Sci. 100, 9383–93387. 494
(doi:10.1073/pnas.1633576100) 495
56. Vázquez DP, Aizen MA. 2004 Asymmetric specialization: a pervasive feature of plant-496
pollinator interactions. Ecology 85, 1251–1257. 497
57. Blüthgen N, Menzel F, Hovestadt T, Fiala B, Blüthgen N. 2007 Specialization, 498
Constraints, and Conflicting Interests in Mutualistic Networks. Curr. Biol. 17, 341–346. 499
(doi:10.1016/j.cub.2006.12.039) 500
58. Balvanera P, Kremen C, Miguel M-R. 2005 Applying community structure analysis to 501
ecosystem function: examples from pollination and carbon storage. Ecol. Appl. 15, 360–502
375. (doi:10.1890/03-5192) 503
59. Kremen C, Williams NM, Thorp RW. 2002 Crop pollination from native bees at risk from 504
agricultural intensification. Proc. Natl. Acad. Sci. 99, 16812–16816. 505
(doi:10.1073/pnas.262413599) 506
60. Klein AM, Steffan-Dewenter I, Tscharntke T. 2003 Fruit set of highland coffee increases 507
with the diversity of pollinating bees. Proc. R. Soc. B Biol. Sci. 270, 955–961. 508
(doi:10.1098/rspb.2002.2306) 509
61. Roswell M, Dushoff J, Winfree R. 2020 Data from: Male and female bees show large 510
25
differences in floral preference. Dryad (doi:https://doi.org/10.5061/dryad.c3rr6q1) 511
62. MacLeod M, Genung MA, Asccher JS, Winfree R. 2016 Measuring partner choice in 512
plant–pollinator networks: using null models to separate rewiring and fidelity from 513
chance. Ecology 97, 2925–2931. 514
63. Roswell M, Dushoff J, Winfree R. 2019 Male and female bees show large differences in 515
floral preference. PLoS One 14, e0214909. (doi:10.1371/journal.pone.0214909) 516
64. Ballantyne G, Baldock KCR, Rendell L, Willmer PG. 2017 Pollinator importance 517
networks illustrate the crucial value of bees in a highly speciose plant community. Sci. 518
Rep. 7, 1–13. (doi:10.1038/s41598-017-08798-x) 519
65. Ne’Eman G, Jürgens A, Newstrom-Lloyd L, Potts SG, Dafni A. 2010 A framework for 520
comparing pollinator performance: Effectiveness and efficiency. Biol. Rev. 85, 435–451. 521
(doi:10.1111/j.1469-185X.2009.00108.x) 522
66. Vázquez DP, Morris WF, Jordano P. 2005 Interaction frequency as a surrogate for the 523
total effect of animal mutualists on plants. Ecol. Lett. 8, 1088–1094. (doi:10.1111/j.1461-524
0248.2005.00810.x) 525
67. Hector A, Bagchi R. 2007 Biodiversity and ecosystem multifunctionality. Nature 448, 526
188–190. (doi:10.1038/nature05947) 527
68. Byrnes JEK et al. 2014 Investigating the relationship between biodiversity and ecosystem 528
multifunctionality: Challenges and solutions. Methods Ecol. Evol. 5, 111–124. 529
(doi:10.1111/2041-210X.12143) 530
69. North BV, Curtis D, Sham PC. 2003 A Note on the Calculation of Empirical P Values 531
26
from Monte Carlo Procedures. Am. J. Hum. Genet. 72, 498–499. (doi:10.1086/346173) 532
70. Rabinowitz D. 1981 Seven forms of rarity. In The Biological Aspects of Rare Plant 533
Conservation (ed H Synge), pp. 205–217. New York: John Wiley & Sons Ltd. 534
71. Loiseau N et al. 2020 Global distribution and conservation status of ecologically rare 535
mammal and bird species. Nat. Commun. 11. (doi:10.1038/s41467-020-18779-w) 536
72. Hercos AP, Sobansky M, Queiroz HL, Magurran AE. 2013 Local and regional rarity in a 537
diverse tropical fish assemblage. Proc. R. Soc. B Biol. Sci. 280. 538
(doi:10.1098/rspb.2012.2076) 539
73. Matias MG, Chapman MG, Underwood AJ, O’Connor NE. 2012 Increasing density of 540
rare species of intertidal gastropods: Tests of competitive ability compared with common 541
species. Mar. Ecol. Prog. Ser. 453, 107–116. (doi:10.3354/meps09639) 542
74. Tilman D, Isbell F, Cowles JM. 2014 Biodiversity and ecosystem functioning. Annu. Rev. 543
Ecol. Syst. 45, 471–493. (doi:10.1146/annurev-ecolsys-120213-091917) 544
75. Leitão RP et al. 2016 Rare species contribute disproportionately to the functional structure 545
of species assemblages. Proc. R. Soc. B 283. (doi:10.1098/rspb.2016.0084) 546
76. Lyons KG, Brigham CA, Traut BH, Schwartz MW. 2005 Rare species and ecosystem 547
functioning. Conserv. Biol. 19, 1019–1024. (doi:10.1111/j.1523-1739.2005.00106.x) 548
77. Dee LE, Cowles J, Isbell F, Pau S, Gaines SD, Reich PB. 2019 When Do Ecosystem 549
Services Depend on Rare Species? Trends Ecol. Evol. 34, 746–758. 550
(doi:10.1016/j.tree.2019.03.010) 551
78. Cane JH, Sipes SS. 2006 Characterizing floral specialization by bees: analytical methods 552
27
and a revised lexicon for oligolecty. In Plant-pollinator interactions: from specialization 553
to generalization (eds NM Waser, J Ollerton), pp. 99–122. Chicago: University of 554
Chicago Press. (doi:10.1007/BF00937854) 555
79. Brosi BJ, Briggs HM. 2013 Single pollinator species losses reduce floral fidelity and plant 556
reproductive function. Proc. Natl. Acad. Sci. 110, 13044–13048. 557
(doi:10.1073/pnas.1307438110) 558
80. CaraDonna PJ, Petry WK, Brennan RM, Cunningham JL, Bronstein JL, Waser NM, 559
Sanders NJ. 2017 Interaction rewiring and the rapid turnover of plant-pollinator networks. 560
Ecol. Lett. 20, 385–394. (doi:10.1111/ele.12740) 561
81. Valdovinos FS, Moisset de Espanés P, Flores JD, Ramos-Jiliberto R. 2013 Adaptive 562
foraging allows the maintenance of biodiversity of pollination networks. Oikos 122, 907–563
917. (doi:10.1111/j.1600-0706.2012.20830.x) 564
82. Hallett AC, Mitchell RJ, Chamberlain ER, Karron JD. 2017 Pollination success following 565
loss of a frequent pollinator: the role of compensatory visitation by other effective 566
pollinators. AoB Plants 9, plx020. 567
83. Morales CL, Traveset A. 2008 Interspecific pollen transfer: Magnitude, prevalence and 568
consequences for plant fitness. CRC. Crit. Rev. Plant Sci. 27, 221–238. 569
(doi:10.1080/07352680802205631) 570
84. Magrach A, Molina FP, Bartomeus I. 2019 Interaction network structure maximizes 571
community-level plant reproduction success via niche complementarity. bioRxiv 572
85. Gonzalez A et al. 2020 Scaling-up biodiversity-ecosystem function research. Ecol. Lett. 573
28
86. Garibaldi LA et al. 2015 Trait matching of flower visitors and crops predicts fruit set 574
better than trait diversity. J. Appl. Ecol. 52, 1436–1444. (doi:10.1111/1365-2664.12530) 575
87. Isbell F et al. 2017 Linking the influence and dependence of people on biodiversity across 576
scales. Nature 546, 65–72. (doi:10.1038/nature22899) 577
88. Isbell F, Cowles J, Dee LE, Loreau M, Reich PB, Gonzalez A, Hector A, Schmid B. 2018 578
Quantifying effects of biodiversity on ecosystem functioning across times and places. 579
Ecol. Lett. 21, 763–778. (doi:10.1111/ele.12928) 580
89. Lefcheck JS et al. 2015 Biodiversity enhances ecosystem multifunctionality across trophic 581
levels and habitats. Nat. Commun. 6, 1–7. (doi:10.1038/ncomms7936) 582
90. Schuldt A et al. 2018 Biodiversity across trophic levels drives multifunctionality in highly 583
diverse forests. Nat. Commun. 9. (doi:10.1038/s41467-018-05421-z) 584
91. Bronstein JL, editor. 2015 Mutualism. Oxford, UK: Oxford University Press. 585
586
Figure legends 587
Figure 1 Hypothetical pollinator abundance distributions illustrating how complementarity 588
among pollinator species should affect the number of functionally important species. (a) The 589
abundance of pollinators visiting the entire plant community. The two most abundant pollinators 590
contribute 80% of floral visits. (b) In a community with low complementarity, the same 591
generalist pollinator species dominate function for every plant species. (c) In a plant community 592
with high complementarity, different pollinator species dominate visits to different plant species. 593
(d) If pollinator species are perfectly redundant (as in b), the number of important pollinator 594
29
species would not change with the number of plant species. If pollinator species are perfectly 595
complementary (similar to c), there would be a positive linear relationship between the number 596
of plants and important pollinators. The real world is likely in between, leading to a positive but 597
saturating relationship. 598
Figure 2 The number of important pollinator species increases with the number of plant species. 599
(a) Accumulation curves for each of the 11 networks. Points represent the number of pollinator 600
species important to at least one plant species in the full community, and lines represent the 601
accumulation of important pollinator species across levels of plant species richness (i.e., means 602
of rarefied plant communities) where the left end represents the average single plant species, and 603
the right end represents the full plant community. (b) An example of one network’s accumulation 604
curve, now shown together with its null model and 95% CIs. The null model curve represents the 605
expectation if individual pollinators forage randomly across the available plant species, while the 606
observed curve includes biological effects, such as species-specific preferences, morphology, or 607
phenology that led to non-random foraging. (c) Z-scores for each network, representing the 608
strength of the biological effects (complementarity) on the number of pollinator species found to 609
be functionally important in a network, relative to the expectation under random foraging. Z-610
scores were calculated as the difference between observed and null expectation (large, red 611
vertical bar in b) divided by the standard deviation of the null (small, blue vertical bar in b) at 612
maximum plant richness for each network (i.e., at the endpoints of the curves in a and b). In (a) 613
and (c), the blue line and points (light grey in black and white) represent the experimental 614
garden. 615
Figure 3 Many rare species are functionally important as pollinators. (a) Pollinator rank 616
abundance distribution for a single network, with pollinator species that were important to at 617
30
least one plant species highlighted in red (appearing darker in black and white). The dotted line 618
represents 1% of total pollinator abundance in the network, which is our definition of rarity. In 619
this community, 13 of 25 important pollinator species are rare. Similar plots for the rest of our 620
communities can be found in the supplement (Figure S7). (b) The proportion of important 621
species that were rare correlated with the number of plant species included in the analysis (r = 622
0.95, p <0.001). When datasets were large enough to include more plant species, more rare 623
pollinator species were found to be important. This increase was only weakly reflected in the null 624
model. The network shown in (a) is circled. 625
Pollinator individuals
d
Pollinator species
a
b
c
0
2
5
5
0
0
1
5
3
0
0
5
0
0
1
0
0
0
0
2
5
5
0
0
1
5
3
0
0
5
0
0
1
0
0
0
Redundancy
Complementarity
Number plant species
Number important
pollinator species
5 10 15 20 25
0 5 15 25 35
5 10 15 20 25
0 5 10 15
0 5 10 15
0 10 20
Observed
Null
a
b
c
Plant species richness
Z
Important pollinator species
5 10 15 20 25
0.0 0.1 0.2 0.3 0.4 0.5
Plant species richness
Proportion of important
species that were rare
Observed
Null
Number of individuals
0 200 600 1000
Supplemental Text
Text S1: Information on datasets
Ten of our 11 datasets were collected from natural or semi-natural meadows of New Jersey. In
these, workers walked transects and collected all bee individuals observed visiting flowers along
that transect.
Six of the meadow datasets were collected by MR and are described in detail in published papers
(Roswell et al. 2019, 2020). In brief, the sites in this study were old fields ranging in size from
0.8 to 2.2 ha, and transects were walked to cover the entire field. Each field was visited multiple
times from June through August, and observations from all visits were combined. These datasets
are large in terms of the number of individual bees collected (mean = 2894.5, sd = 1091.1) and
the number of plant species included (mean = 17.7, sd = 4.3).
Four meadow datasets are published for the first time here. These data were collected in natural
or semi-natural meadows of New Jersey, USA, in 2011, 2013, and 2014, as part of a larger,
unpublished study. Each study site included paired, adjacent fields, one of which was
unmanipulated and one of which had been seeded with wild flowers. In each of the four study
sites included here, eight 40x2 m transects (four in each field) were positioned within 250 m of
one another. Bees were collected along transects in 10-min sampling bouts in which a data
collector slowly walked the transect (10 m per 2.5 min) and collected all bees observed on
flowers within 1 m of the transect. Each site was visited four to seven times from May through
August or September, and each site visit included 2-4 sampling bouts. Due to less intensive
sampling effort, these datasets are smaller in terms of the number of individual bees collected
(302.3, sd = 64.1) and the number of plant species included (mean = 6.25, sd = 0.5).
The final dataset included in this analysis is wild bee visits to an outdoor experimental garden,
which is described in detail in a published paper (MacLeod et al 2016). The garden was planted
in a 6 x 17 block design, in which 17 plant species were planted in 1 m2 monoculture plots,
randomly arranged in each of 6 blocks (i.e. 6 rows), such that each plant species was replicated
six times. Total garden area was ~1600 m2. Bees were collected during repeated, timed
observations of each plot. Thus, the abundance of each plant species was standardized by area,
and sampling effort was standardized across plant species. This dataset was relatively large,
including 2367 individual bees, and 17 plant species.
Text S2: Identifying “functionally important bee species”
In this study, we define “important” bee species as those providing some threshold proportion of
visits to a given plant species. In doing so, our aim is to identify the bee species that account for
a substantial proportion of all bee visits to a given plant species. We acknowledge that any
choice of threshold is to some extent arbitrary; we are collapsing a continuous measure to a
binary measure. In the analysis of real-world ecosystem functions, however, it is often necessary
to use some method to distinguish the more functionally important species, to avoid counting all
species, including those that may have been observed only once or twice in a large study, as
essential to function. Similarly, real-world datasets will always be subject to sampling effects
(e.g., increases in the number of species recorded with increased sampling effort) because it is
not possible to sample natural communities completely. Focusing on those species that account
for a large(r) proportion of individuals collected will mitigate the impact of sampling effects on
the analysis, because the more common species are better sampled than the rare ones for a given
level of sampling effort.
In the main text, we use a 5% threshold, such that any bee species contributing at least 5% of the
visits to a given plant species is considered important. We believe this captures 'functionally
important' species in a biologically reasonable way. When using the 5% threshold, the important
species provide 81% of the bee visits per plant species on average (sd = 8.9%), while
representing only 35% of the bee species per plant species (sd = 18%). Fig. S2 shows what the
5% threshold means graphically by plotting example rank-abundance distributions of bees per
plant species, with the functionally important bee species shaded. Fig. S3 shows how the
proportion of species considered important and the visits they provide changes with the choice of
threshold.
Text S3: Sensitivity analyses
We re-ran each of our analyses using a range of thresholds to define bee importance, from bees
providing at least 2.5% of a plant’s visits to those providing at least 10% of visits, by increments
of 0.1% (76 total runs). Because the null model takes hours to run, however, we only ran the null
model four times, for thresholds of 2.5, 5, 7.5, and 10%. For plant species with few observations,
low importance thresholds can mean that every bee species is considered important, even bee
species with only one visit. To avoid this, we included a condition that a bee must meet the
threshold for importance and be represented by >1 individual.
A lower, more inclusive threshold would mean that more bee species would be considered
functionally important to each plant, raising the intercepts of the accumulation curves. A higher,
less inclusive threshold should result in the opposite. Having a larger number of important
species also creates more opportunity for turnover of important species, so we might also expect
the lower, more inclusive thresholds to result in steeper accumulation curves. Indeed, the
absolute number of important bee species varied with the choice of threshold as expected (Figs
S5a, S6). Yet, the factor-increase in the total number of functionally important bees between a
single plant species and its community varied little, ranging from 4.3 to 5.2, and actually
increasing with the threshold used (Figs S5b, S6). Z-scores also increased with higher
thresholds, from an average of 5.7 at the more inclusive threshold of 2.5%, to 8.1 at the less
inclusive threshold of 10% (Fig S6). This also meant that, at higher thresholds, more networks
were significantly different than the null; at the 2.5% threshold, none of the four smallest
networks significantly differed from the null, while at the 10% threshold all networks
significantly differed from the null. This is because the null models were flatter and less variable
under higher thresholds, accentuating differences between the observed results and null
expectation. In turn, this is likely because of the skewed bee species abundance distributions; at
higher thresholds, when only highly dominant bee species are considered important, it is more
difficult for less abundant bees to appear important by chance. As a result, there is less turnover
of important bees across plants, and less variation in results between runs of the null model. In
sum, while the number of bee species defined as 'important' changed with the threshold used (as
would have to be the case), the main results of the study (the factor increase across the species
accumulation curve, and the Z score representing the biological component of the increase) were
qualitatively similar across thresholds.
Supplemental figures
Fig S1 Map of our study plots in New Jersey, USA. The experimental garden is the southern-
most plot.
Fig S2 Example rank abundance distributions of bee species visiting four different plant species,
with important bees (those contributing ≥5% of visits) shown in red. These distributions
demonstrate the range of what the 5% importance thresholds look like in practice. Most plant
species’ distributions look like (a) or (b), in which relatively few bee species are considered
important. For some plants, however, quirks of small sample sizes or more even bee abundance
distributions led to a high proportion bee species being considered important, as in (c) and (d).
Fig. S3 Effect of the threshold used to define importance on the proportion of bee species
considered important and the proportion of visits provided by those important species. (a-b) The
proportion of species considered important declines with higher importance thresholds; (a)
shows the mean, across networks, and (b) shows the distribution of values across networks. (c-d)
Because fewer species are considered important with higher thresholds, the proportion of visits
contributed by those species also decreases. However, across all thresholds we considered (2.5-
10% of visits), important bee species account for a high proportion of total flower visits (ca. 70-
90%).
Figure S4. Observed and null accumulation curves for each network. Text in the upper left of
each panel refers to the network (see Table S1); ‘cm13’ is the experimental garden and the rest
are natural / semi-natural meadow communities. Each curve describes the mean number of
important bee species, taken across combinations of plant species. The ribbon around the
observed accumulation curve shows the inner 95th percentile (i.e., 2.5th to 97.5th percentile) at
each rarefied value of plant richness, and thus describes variation in the number of important bee
species that is due to plant identity. The vertical bars on the null accumulation curve were
similarly calculated, but represent the inner 95th percentile of means across iterations of the null
model. This variation comes from stochastic sampling effects under an assumption of no true
differences among bee species in the plant species they visit. Thus, these vertical bars represent
95% confidence intervals around the mean expectation of the null model. Comparing the mean
observation (the black curve) to the null expectation and its CI (grey curve and vertical bars) asks
whether, on average, more bee species are important than expected due to sampling effects
alone. In 9 of our 11 networks, the answer is yes.
Fig S5a (cont’d next page)
Figure S5 Sensitivity analysis showing the accumulation of important bee species with
increasing plant species richness under varying thresholds of importance. Text in the upper left
of each panel refers to the network/dataset (see Table S1); ‘cm13’ is the experimental garden and
the rest are natural / semi-natural meadow communities. We re-ran our analysis while defining
important bees as those contributing 2.5% to 10% of visits to at least one plant species, using
increments of 0.1%. (a) Increasing the threshold of importance lowers the absolute number of
important species. (b) However, the proportional change in the number important bee species,
relative to a single plant species, was very robust to changes in threshold definition, and actually
tended to increase with higher thresholds.
Fig S6 Effect of the threshold of importance on our main results. Each column is analogous to
the main text Fig. 2. The first row shows the accumulation curves for each network. Points
represent the number of bee species important to at least one plant species in the full community,
and grey lines represent the accumulation of important species across rarefied levels of plant
species richness (i.e., means of across rarefied plant communities). The r value refers to the
correlation between plant species richness and the number of important bees in the whole
network (i.e. the black points). The second row shows the accumulation curve of a typical
network together with its null model. The third row shows Z-scores for the number of important
bees in each network, and r values refer to the correlation between these Z-scores and plant
species richness. Although the absolute number of important bee species decreased with
increasing thresholds, the relationship between important bees and plant richness remains largely
unchanged. In fact, the average Z-score goes up under higher thresholds because the null models
get flatter and less variable.
Figure S7 Log-transformed rank abundance distributions for each of the networks in our
analysis with the important bee species (those contributing at least 5% of total bee visits to at
least one plant species) shown in red. The horizontal dashed line represents 1% of the total bee
community; bee species below this line are considered locally rare. Text in the upper right of
each panel refers to the network/dataset (see Table S1); ‘cm13’ is the experimental garden and
the rest are natural / semi-natural meadow communities. In most plant communities, many rare
bees are also important to function of at least one plant species.
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Fig S8 (cont’d next page)
Figure S8. Bee abundances by rank in each of our networks (large insets at the top of each page)
and on each plant species in each network (small insets). These plots are analogous to Figure 1 a-
c. In the network-wide abundance distributions (large insets), dotted horizontal lines denote 5%
and 1% of the total bee community. In the plant-level abundance distributions (small insets),
horizontal line denotes 5% of the individuals visiting that plant species. The insets for each plant
species are arranged in descending order (left to right, top to bottom) by the number of
observations on each plant species. These figures are interpreted as in Figure 1: if bee species are
mostly redundant, the same dominant bee species (in blue) will tend to dominate visits to
individual plant species. Complementarity among bee species, on the other hand, will lead to
subdominant or even rare species (in grey and red) to dominate visits to some plant species.
Figure S9 Sensitivity analysis showing the proportion of important bee species that are rare,
across networks, using different rarity thresholds. In this figure, we held the threshold of
importance at 5% of visits and used rarity thresholds of 0.5% to 1.5%, by increments of 0.01%.
Raising the rarity threshold means more important bee species are considered rare, and vice
versa. In the main text, we used a threshold of 1%. However, regardless of the threshold, our
conclusion remains the same: across networks, a substantial proportion of important bee species
are also locally rare.
Figure S10 Sensitivity analysis showing the proportion of important bees that were rare using
different thresholds of both bee importance and rarity. The value shown is the mean proportion
across all 11 networks. The proportion of important bees that are rare increases with lower
thresholds of importance and higher thresholds of rarity. Within this parameter space, the mean
(across networks) proportion of important bees that are considered rare ranges from 7.8% in the
bottom right to 44.8% in the top left. In the main text, we used thresholds of 5% for importance
and 1.0% for rarity, resulting in a mean of 25.0% of important bees being considered rare (shown
as white). Although this value is dependent on the thresholds used, our general conclusion holds
across networks: it is common for important bees to also be rare, even under strict definitions of
importance and rarity.
Supplemental tables
Table S1. Datasets, or networks, included in our analysis. Each network was collected at one site in one year. Cape May (cm13) is the
experimental garden, in which plant abundance was standardized (by area) ad sampling effort was standardized across plant species. The column
Unique Sampling Days refers to the number of different days during the focal year on which sampling occurred. Total Sample Effort refers to the
total number of minutes spent netting, summed across days and data collectors. In the remaining columns, we include values for the entire dataset
as collected, and the subset of the data we analyzed here (by including only plant species with ≥ 20 observed interactions).
Network
Citation
Unique
sample
dates
Total sample
effort
(minutes)
Plant sp
(original/subset)
Bee sp
(original/subset)
Observations
(original/subset)
Unique
interactions
(original/subset)
Cm13
MacLeod et al 2016
30
5280
17/17
54/54
2346/2346
243/243
Baldpate
Roswell et al 2019
14
2400
47/21
69/68
4629/4513
383/313
Cold Soil
Roswell et al 2019
13
2580
52/23
84/83
2637/2498
447/375
Fox Hill
Roswell et al 2019
16
3690
37/20
74/70
3714/3582
363/306
IAS
Roswell et al 2019
16
3810
36/16
88/86
3342/3209
380/310
Lord Stirling
Roswell et al 2019
14
3690
26/13
65/64
2098/2039
268/232
URWA
Roswell et al 2019
16
4020
26/13
74/71
1594/1526
254/211
BW_2014
This paper
4
80
14/6
35/30
334/271
91/59
HP_2011
This paper
7
250
21/7
33/29
415/353
109/66
MU_2013
This paper
6
160
36/6
42/37
488/358
133/62
RO_2014
This paper
4
80
22/6
33/22
289/227
80/40
References
MacLeod, M., Genung, M.A., Asccher, J.S. & Winfree, R. (2016). Measuring partner choice in plant–pollinator networks: using null
models to separate rewiring and fidelity from chance. Ecology, 97, 2925–2931.
Roswell, M., Dushoff, J. & Winfree, R. (2019). Male and female bees show large differences in floral preference. PLoS One, 14,
e0214909.
Roswell, M.; Dushoff, J. & Winfree, R. (2020), Data from: Male and female bees show large differences in floral preference, v3,
Dryad, Dataset, https://doi.org/10.5061/dryad.c3rr6q1
Table S2. Plant species and the number of individual bees observed on that species, for each network as
analyzed.
Network
Plant species
Observed
interactions
Baldpate
Monarda fistulosa
895
Baldpate
Rudbeckia hirta
866
Baldpate
Pycnanthemum muticum
452
Baldpate
Trifolium campestre
336
Baldpate
Erechtites hieraciifolius
328
Baldpate
Linaria vulgaris
271
Baldpate
Daucus carota
255
Baldpate
Phytolacca americana
181
Baldpate
Nepeta cataria
166
Baldpate
Asclepias tuberosa
129
Baldpate
Verbascum thapsus
108
Baldpate
Hypericum perforatum
104
Baldpate
Erigeron strigosus
91
Baldpate
Cirsium vulgare
69
Baldpate
Trifolium repens
61
Baldpate
Pycnanthemum tenuifolium
47
Baldpate
Verbena urticifolia
42
Baldpate
Penstemon hirsutus
35
Baldpate
Conyza canadensis
33
Baldpate
Potentilla recta
23
Baldpate
Plantago lanceolata
21
BW_2014
Lespedeza cuneata
76
BW_2014
Monarda fistulosa
54
BW_2014
Erigeron annuus
39
BW_2014
Symphyotrichum novae-angliae
39
BW_2014
Echinacea purpurea
34
BW_2014
Gaillardia pulchella
29
cm13
Pycnanthemum tenuifolium
797
cm13
Symphyotrichum novae-angliae
275
cm13
Eupatoriadelphus maculatus
233
cm13
Asclepias tuberosa
199
cm13
Rudbeckia laciniata
135
cm13
Veronicastrum virginicum
110
cm13
Oligoneuron rigidum
88
cm13
Penstemon hirsutus
83
cm13
Vernonia noveboracensis
66
cm13
Asclepias incarnata
62
cm13
Rudbeckia hirta
57
cm13
Lobelia siphilitica
50
cm13
Solidago rugosa
50
cm13
Verbena hastata
50
cm13
Zizia aurea
41
cm13
Euthamia graminifolia
37
cm13
Agastache scrophulariifolia
34
Cold Soil
Lythrum salicaria
303
Cold Soil
Monarda fistulosa
277
Cold Soil
Pycnanthemum tenuifolium
256
Cold Soil
Daucus carota
254
Cold Soil
Chamaecrista fasciculata
211
Cold Soil
Erigeron strigosus
177
Cold Soil
Rudbeckia hirta
123
Cold Soil
Solidago juncea
122
Cold Soil
Solidago gigantea
108
Cold Soil
Trifolium pratense
82
Cold Soil
Heliopsis helianthoides
81
Cold Soil
Plantago lanceolata
79
Cold Soil
Achillea millefolium
59
Cold Soil
Vernonia noveboracensis
54
Cold Soil
Echinacea purpurea
49
Cold Soil
Leucanthemum vulgare
48
Cold Soil
Penstemon digitalis
44
Cold Soil
Euthamia graminifolia
36
Cold Soil
Apocynum cannabinum
33
Cold Soil
Melilotus albus
29
Cold Soil
Erigeron annuus
28
Cold Soil
Trifolium repens
23
Cold Soil
Cirsium arvense
22
Fox Hill
Daucus carota
900
Fox Hill
Ratibida pinnata
660
Fox Hill
Erigeron strigosus
375
Fox Hill
Monarda fistulosa
322
Fox Hill
Leucanthemum vulgare
285
Fox Hill
Cirsium arvense
141
Fox Hill
Rudbeckia hirta
110
Fox Hill
Asclepias syriaca
107
Fox Hill
Cichorium intybus
106
Fox Hill
Verbena urticifolia
84
Fox Hill
Centaurea stoebe
78
Fox Hill
Echinacea purpurea
76
Fox Hill
Melilotus officinalis
62
Fox Hill
Cirsium vulgare
59
Fox Hill
Clinopodium vulgare
45
Fox Hill
Solidago juncea
44
Fox Hill
Penstemon digitalis
36
Fox Hill
Apocynum cannabinum
33
Fox Hill
Trifolium repens
32
Fox Hill
Potentilla recta
27
HP_2011
Clinopodium vulgare
111
HP_2011
Echinacea purpurea
103
HP_2011
Erigeron philadelphicus
34
HP_2011
Pastinaca sativa
31
HP_2011
Solidago gigantea
31
HP_2011
Daucus carota
22
HP_2011
Solidago canadensis
21
IAS
Pycnanthemum tenuifolium
689
IAS
Daucus carota
515
IAS
Solidago juncea
380
IAS
Penstemon digitalis
323
IAS
Liatris spicata
311
IAS
Achillea millefolium
264
IAS
Heliopsis helianthoides
139
IAS
Rudbeckia hirta
124
IAS
Trifolium pratense
119
IAS
Hypericum perforatum
100
IAS
Monarda fistulosa
96
IAS
Tradescantia ohiensis
34
IAS
Symphyotrichum novae-angliae
31
IAS
Cirsium arvense
30
IAS
Drymocallis arguta
30
IAS
Baptisia tinctoria
24
Lord Stirling
Eutrochium maculatum
554
Lord Stirling
Penstemon digitalis
382
Lord Stirling
Cirsium arvense
254
Lord Stirling
Apocynum cannabinum
196
Lord Stirling
Monarda fistulosa
154
Lord Stirling
Pycnanthemum tenuifolium
140
Lord Stirling
Achillea millefolium
137
Lord Stirling
Lythrum salicaria
51
Lord Stirling
Lotus corniculatus
43
Lord Stirling
Securigera varia
39
Lord Stirling
Daucus carota
36
Lord Stirling
Rudbeckia hirta
32
Lord Stirling
Solidago juncea
21
MU_2013
Solidago gigantea
140
MU_2013
Monarda fistulosa
65
MU_2013
Cirsium arvense
57
MU_2013
Galium palustre
42
MU_2013
Lythrum salicaria
31
MU_2013
Coreopsis lanceolata
23
RO_2014
Centaurea maculosa
54
RO_2014
Leucanthemum vulgare
51
RO_2014
Solidago canadensis
35
RO_2014
Euthamia graminifolia
34
RO_2014
Monarda fistulosa
31
RO_2014
Rudbeckia laciniata
22
URWA
Pycnanthemum tenuifolium
356
URWA
Centaurea stoebe
287
URWA
Solidago juncea
137
URWA
Erigeron strigosus
136
URWA
Penstemon digitalis
115
URWA
Daucus carota
103
URWA
Lotus corniculatus
92
URWA
Apocynum cannabinum
87
URWA
Leucanthemum vulgare
78
URWA
Rosa carolina
47
URWA
Rudbeckia hirta
45
URWA
Achillea millefolium
23
URWA
Clinopodium vulgare
20
Table S3. Bee species and the number of plant visits by that species observed in the dataset as analyzed.
Colors are as in Figure S8: Species that were dominant in their community (> 5% of observations) are in
blue, rare species (< 1% of observations) are in red, and subdominants (1-5% of observations) are in grey.
Bee species that were important to at least one plant species in its network (i.e., providing >5% of visits to
that plant species) are bolded.
Network
Bee species
Observations
Baldpate
Augochlora pura
1263
Baldpate
Bombus impatiens
935
Baldpate
Bombus griseocollis
387
Baldpate
Ceratina calcarata
380
Baldpate
Bombus bimaculatus
209
Baldpate
Halictus ligatus
209
Baldpate
Hylaeus affinis_modestus
180
Baldpate
Andrena wilkella
150
Baldpate
Augochloropsis metallica
120
Baldpate
Augochlorella aurata
105
Baldpate
Ceratina strenua
69
Baldpate
Lasioglossum imitatum
65
Baldpate
Lasioglossum versatum
65
Baldpate
Lasioglossum hitchensi_weemsi
33
Baldpate
Agapostemon virescens
31
Baldpate
Halictus confusus
30
Baldpate
Megachile mendica
22
Baldpate
Ceratina dupla
21
Baldpate
Halictus parallelus
18
Baldpate
Megachile exilis
18
Baldpate
Lasioglossum illinoense
17
Baldpate
Hylaeus mesillae
16
Baldpate
Xylocopa virginica
12
Baldpate
Lasioglossum cressonii
11
Baldpate
Lasioglossum cattellae
10
Baldpate
Osmia bucephala
10
Baldpate
Lasioglossum tegulare
9
Baldpate
Ceratina mikmaqi
8
Baldpate
Sphecodes heraclei
8
Baldpate
Bombus perplexus
7
Baldpate
Megachile campanulae
7
Baldpate
Halictus rubicundus
6
Baldpate
Lasioglossum coriaceum
6
Baldpate
Bombus citrinus
5
Baldpate
Bombus fervidus
5
Baldpate
Lasioglossum trigeminum
5
Baldpate
Heriades carinata
4
Baldpate
Lasioglossum abanci
4
Baldpate
Lasioglossum fuscipenne
4
Baldpate
Hoplitis pilosifrons
3
Baldpate
Hoplitis producta
3
Baldpate
Lasioglossum subviridatum
3
Baldpate
Osmia pumila
3
Baldpate
Andrena rudbeckiae
2
Baldpate
Bombus vagans
2
Baldpate
Calliopsis andreniformis
2
Baldpate
Lasioglossum ellisiae
2
Baldpate
Lasioglossum pectorale
2
Baldpate
Lithurgus chrysurus
2
Baldpate
Megachile petulans
2
Baldpate
Megachile rotundata
2
Baldpate
Megachile xylocopoides
2
Baldpate
Melissodes desponsus
2
Baldpate
Melissodes subillatus
2
Baldpate
Osmia atriventris
2
Baldpate
Agapostemon sericeus
1
Baldpate
Anthophora abrupta
1
Baldpate
Bombus auricomus
1
Baldpate
Coelioxys octodentatus
1
Baldpate
Coelioxys sayi
1
Baldpate
Hoplitis spoliata
1
Baldpate
Lasioglossum callidum
1
Baldpate
Lasioglossum coeruleum
1
Baldpate
Lasioglossum foxii
1
Baldpate
Lasioglossum rozeni
1
Baldpate
Megachile frugalis
1
Baldpate
Megachile pugnata
1
Baldpate
Melissodes denticulatus
1
BW_2014
Bombus impatiens
65
BW_2014
Bombus griseocollis
38
BW_2014
Ceratina calcarata
36
BW_2014
Halictus ligatus
18
BW_2014
Megachile mendica
15
BW_2014
Bombus bimaculatus
14
BW_2014
Augochlora pura
11
BW_2014
Hylaeus mesillae
9
BW_2014
Lasioglossum pilosum
8
BW_2014
Xylocopa virginica
8
BW_2014
Lasioglossum hitchensi_weemsi
7
BW_2014
Ceratina strenua
6
BW_2014
Hylaeus affinis_modestus
5
BW_2014
Lasioglossum trigeminum
5
BW_2014
Lasioglossum imitatum
4
BW_2014
Megachile inimica
3
BW_2014
Agapostemon texanus
2
BW_2014
Agapostemon virescens
2
BW_2014
Megachile brevis
2
BW_2014
Megachile exilis
2
BW_2014
Megachile pugnata
2
BW_2014
Bombus fervidus
1
BW_2014
Bombus perplexus
1
BW_2014
Coelioxys sayi
1
BW_2014
Heriades variolosa
1
BW_2014
Lasioglossum leucocomum
1
BW_2014
Lasioglossum pectorale
1
BW_2014
Lasioglossum platyparium
1
BW_2014
Lasioglossum versatum
1
BW_2014
Stelis lateralis
1
cm13
Bombus griseocollis
429
cm13
Bombus impatiens
340
cm13
Lasioglossum vierecki
246
cm13
Agapostemon virescens
173
cm13
Lasioglossum leucocomum
168
cm13
Megachile mendica
136
cm13
Halictus ligatus_poeyi
128
cm13
Agapostemon texanus
66
cm13
Xylocopa virginica
57
cm13
Ceratina mikmaqi
53
cm13
Hoplitis pilosifrons
53
cm13
Lasioglossum pilosum
44
cm13
Hylaeus mesillae
39
cm13
Bombus bimaculatus
37
cm13
Halictus confusus
36
cm13
Augochloropsis metallica
31
cm13
Lasioglossum tegulare
31
cm13
Bombus pensylvanicus
30
cm13
Sphecodes cressonii
24
cm13
Augochlorella aurata
21
cm13
Hylaeus affinis_modestus
21
cm13
Lasioglossum pectorale
21
cm13
Ceratina calcarata
20
cm13
Apis mellifera
18
cm13
Coelioxys sayi
13
cm13
Coelioxys octodentatus
11
cm13
Epeolus lectoides
11
cm13
Melissodes subillatus
11
cm13
Megachile brevis
10
cm13
Megachile inimica
10
cm13
Augochlora pura
9
cm13
Lasioglossum hitchensi
9
cm13
Ceratina dupla
7
cm13
Colletes simulans
6
cm13
Megachile exilis
6
cm13
Megachile xylocopoides
5
cm13
Sphecodes atlantis
5
cm13
Nomada vegana
4
cm13
Andrena atlantica
3
cm13
Heriades leavitti
3
cm13
Megachile texana
3
cm13
Agapostemon splendens
2
cm13
Ceratina strenua
2
cm13
Lasioglossum coreopsis
2
cm13
Megachile campanulae
2
cm13
Stelis louisae
2
cm13
Andrena nasonii
1
cm13
Bombus fervidus
1
cm13
Eucera hamata
1
cm13
Megachile gemula
1
cm13
Megachile rotundata
1
cm13
Melissodes trinodis
1
cm13
Nomada affabilis
1
cm13
Nomada australis
1
cm13
Sphecodes mandibularis
1
Cold Soil
Bombus impatiens
607
Cold Soil
Halictus ligatus
402
Cold Soil
Bombus griseocollis
191
Cold Soil
Lasioglossum imitatum
150
Cold Soil
Hylaeus affinis_modestus
116
Cold Soil
Lasioglossum versatum
107
Cold Soil
Hylaeus mesillae
80
Cold Soil
Halictus confusus
74
Cold Soil
Ceratina calcarata
67
Cold Soil
Xylocopa virginica
56
Cold Soil
Ceratina strenua
55
Cold Soil
Lasioglossum oceanicum
50
Cold Soil
Andrena wilkella
41
Cold Soil
Bombus bimaculatus
41
Cold Soil
Ceratina mikmaqi
40
Cold Soil
Ceratina dupla
36
Cold Soil
Melissodes denticulatus
28
Cold Soil
Lasioglossum illinoense
25
Cold Soil
Augochlora pura
22
Cold Soil
Lasioglossum hitchensi_weemsi
22
Cold Soil
Melissodes subillatus
19
Cold Soil
Augochlorella aurata
17
Cold Soil
Megachile sculpturalis
17
Cold Soil
Andrena rudbeckiae
16
Cold Soil
Megachile mendica
16
Cold Soil
Halictus rubicundus
12
Cold Soil
Megachile montivaga
12
Cold Soil
Augochloropsis metallica
11
Cold Soil
Agapostemon virescens
10
Cold Soil
Hoplitis pilosifrons
10
Cold Soil
Megachile brevis
10
Cold Soil
Lasioglossum trigeminum
9
Cold Soil
Lasioglossum viridatum
9
Cold Soil
Megachile frugalis
9
Cold Soil
Bombus fervidus
6
Cold Soil
Heriades carinata
6
Cold Soil
Melissodes agilis
6
Cold Soil
Anthidium manicatum
5
Cold Soil
Lasioglossum zephyrum
5
Cold Soil
Megachile exilis
5
Cold Soil
Megachile rotundata
5
Cold Soil
Melissodes bimaculatus
5
Cold Soil
Sphecodes dichrous
5
Cold Soil
Anthidium oblongatum
4
Cold Soil
Lasioglossum admirandum
4
Cold Soil
Lasioglossum bruneri
4
Cold Soil
Lasioglossum callidum
4
Cold Soil
Anthidiellum notatum
3
Cold Soil
Bombus perplexus
3
Cold Soil
Lasioglossum coriaceum
3
Cold Soil
Megachile pugnata
3
Cold Soil
Hylaeus leptocephalus
2
Cold Soil
Lasioglossum leucozonium
2
Cold Soil
Megachile campanulae
2
Cold Soil
Bombus vagans
1
Cold Soil
Calliopsis andreniformis
1
Cold Soil
Coelioxys alternatus
1
Cold Soil
Coelioxys germanus
1
Cold Soil
Coelioxys hunteri
1
Cold Soil
Coelioxys obtusiventris
1
Cold Soil
Coelioxys octodentatus
1
Cold Soil
Heriades variolosa
1
Cold Soil
Hoplitis producta
1
Cold Soil
Lasioglossum atwoodi
1
Cold Soil
Lasioglossum cressonii
1
Cold Soil
Lasioglossum gotham
1
Cold Soil
Lasioglossum obscurum
1
Cold Soil
Lasioglossum paradmirandum
1
Cold Soil
Lasioglossum pilosum
1
Cold Soil
Lasioglossum rozeni
1
Cold Soil
Lasioglossum smilacinae
1
Cold Soil
Lasioglossum tegulare
1
Cold Soil
Megachile integra
1
Cold Soil
Melissodes desponsus
1
Cold Soil
Melissodes trinodis
1
Cold Soil
Nomada erigeronis
1
Cold Soil
Pseudoanthidium nanum
1
Cold Soil
Sphecodes heraclei
1
Cold Soil
Stelis lateralis
1
Cold Soil
Stelis louisae
1
Cold Soil
Triepeolus cressonii
1
Cold Soil
Triepeolus eliseae
1
Cold Soil
Triepeolus remigatus
1
Fox Hill
Halictus ligatus
760
Fox Hill
Augochlorella persimilis
515
Fox Hill
Hylaeus affinis_modestus
498
Fox Hill
Bombus griseocollis
265
Fox Hill
Ceratina calcarata
264
Fox Hill
Hylaeus mesillae
206
Fox Hill
Lasioglossum versatum
203
Fox Hill
Augochlorella aurata
132
Fox Hill
Bombus impatiens
128
Fox Hill
Lasioglossum imitatum
89
Fox Hill
Bombus bimaculatus
51
Fox Hill
Agapostemon virescens
49
Fox Hill
Lasioglossum hitchensi_weemsi
38
Fox Hill
Bombus perplexus
29
Fox Hill
Augochlora pura
26
Fox Hill
Ceratina dupla
26
Fox Hill
Lasioglossum oceanicum
25
Fox Hill
Ceratina strenua
24
Fox Hill
Lasioglossum gotham
23
Fox Hill
Halictus confusus
17
Fox Hill
Lasioglossum rozeni
17
Fox Hill
Andrena wilkella
15
Fox Hill
Ceratina mikmaqi
15
Fox Hill
Hoplitis pilosifrons
15
Fox Hill
Lasioglossum trigeminum
15
Fox Hill
Halictus rubicundus
9
Fox Hill
Lasioglossum tegulare
9
Fox Hill
Megachile mendica
9
Fox Hill
Heriades carinata
8
Fox Hill
Augochloropsis metallica
7
Fox Hill
Lasioglossum admirandum
7
Fox Hill
Lasioglossum cressonii
6
Fox Hill
Bombus vagans
5
Fox Hill
Megachile brevis
5
Fox Hill
Nomada pygmaea
5
Fox Hill
Andrena nasonii
4
Fox Hill
Lasioglossum cattellae
4
Fox Hill
Lasioglossum platyparium
4
Fox Hill
Osmia bucephala
4
Fox Hill
Hoplitis producta
3
Fox Hill
Lasioglossum illinoense
3
Fox Hill
Melissodes subillatus
3
Fox Hill
Osmia atriventris
3
Fox Hill
Osmia georgica
3
Fox Hill
Osmia pumila
3
Fox Hill
Xylocopa virginica
3
Fox Hill
Andrena pruni
2
Fox Hill
Lasioglossum callidum
2
Fox Hill
Lasioglossum coriaceum
2
Fox Hill
Megachile montivaga
2
Fox Hill
Melissodes bimaculatus
2
Fox Hill
Nomada maculata
2
Fox Hill
Andrena brevipalpis
1
Fox Hill
Andrena commoda
1
Fox Hill
Andrena imitatrix
1
Fox Hill
Anthidium manicatum
1
Fox Hill
Heriades variolosa
1
Fox Hill
Hoplitis spoliata
1
Fox Hill
Lasioglossum abanci
1
Fox Hill
Lasioglossum ellisiae
1
Fox Hill
Lasioglossum foxii
1
Fox Hill
Lasioglossum nigroviride
1
Fox Hill
Lasioglossum paradmirandum
1
Fox Hill
Lasioglossum smilacinae
1
Fox Hill
Lasioglossum zephyrum
1
Fox Hill
Megachile campanulae
1
Fox Hill
Megachile exilis
1
Fox Hill
Megachile sculpturalis
1
Fox Hill
Melissodes denticulatus
1
Fox Hill
Melissodes desponsus
1
HP_2011
Bombus impatiens
145
HP_2011
Bombus griseocollis
24
HP_2011
Halictus ligatus
24
HP_2011
Lasioglossum imitatum
23
HP_2011
Hylaeus affinis_modestus
19
HP_2011
Hylaeus mesillae
18
HP_2011
Ceratina dupla
16
HP_2011
Ceratina mikmaqi
13
HP_2011
Ceratina strenua
11
HP_2011
Lasioglossum hitchensi_weemsi
9
HP_2011
Agapostemon virescens
7
HP_2011
Ceratina calcarata
7
HP_2011
Andrena brevipalpis
6
HP_2011
Lasioglossum versatum
5
HP_2011
Xylocopa virginica
5
HP_2011
Augochlorella aurata
4
HP_2011
Lasioglossum pectorale
4
HP_2011
Melissodes bimaculatus
2
HP_2011
Andrena hirticincta
1
HP_2011
Andrena nasonii
1
HP_2011
Andrena simplex
1
HP_2011
Anthidium manicatum
1
HP_2011
Augochlora pura
1
HP_2011
Colletes compactus
1
HP_2011
Colletes simulans
1
HP_2011
Hoplitis pilosifrons
1
HP_2011
Lasioglossum obscurum
1
HP_2011
Lasioglossum trigeminum
1
HP_2011
Pseudopanurgus andrenoides
1
IAS
Halictus ligatus
689
IAS
Hylaeus affinis_modestus
520
IAS
Bombus griseocollis
284
IAS
Bombus impatiens
277
IAS
Ceratina calcarata
216
IAS
Bombus bimaculatus
211
IAS
Hylaeus mesillae
191
IAS
Agapostemon virescens
135
IAS
Lasioglossum versatum
88
IAS
Lasioglossum oceanicum
46
IAS
Lasioglossum imitatum
44
IAS
Augochlorella aurata
43
IAS
Lasioglossum callidum
42
IAS
Andrena wilkella
38
IAS
Halictus confusus
32
IAS
Ceratina strenua
28
IAS
Xylocopa virginica
23
IAS
Megachile mendica
15
IAS
Hoplitis pilosifrons
14
IAS
Lasioglossum illinoense
14
IAS
Megachile exilis
13
IAS
Halictus rubicundus
11
IAS
Lasioglossum hitchensi_weemsi
11
IAS
Megachile sculpturalis
11
IAS
Heriades carinata
10
IAS
Lasioglossum gotham
10
IAS
Augochloropsis metallica
9
IAS
Ceratina dupla
9
IAS
Ceratina mikmaqi
9
IAS
Heriades variolosa
8
IAS
Megachile frugalis
8
IAS
Augochlora pura
7
IAS
Megachile campanulae
7
IAS
Megachile rotundata
7
IAS
Nomada pygmaea
7
IAS
Agapostemon sericeus
6
IAS
Bombus fervidus
6
IAS
Lasioglossum rozeni
6
IAS
Nomada bidentate_group
6
IAS
Anthidium oblongatum
5
IAS
Heriades leavitti
5
IAS
Lasioglossum trigeminum
5
IAS
Melissodes subillatus
5
IAS
Melissodes trinodis
5
IAS
Andrena nasonii
4
IAS
Halictus parallelus
4
IAS
Megachile inimica
4
IAS
Osmia bucephala
4
IAS
Sphecodes heraclei
4
IAS
Andrena carlini
3
IAS
Anthophora terminalis
3
IAS
Lasioglossum cattellae
3
IAS
Osmia distincta
3
IAS
Sphecodes dichrous
3
IAS
Lasioglossum admirandum
2
IAS
Lasioglossum tegulare
2
IAS
Lasioglossum zephyrum
2
IAS
Megachile montivaga
2
IAS
Nomada articulata
2
IAS
Osmia pumila
2
IAS
Andrena cressonii
1
IAS
Andrena spiraeana
1
IAS
Andrena vicina
1
IAS
Anthidiellum notatum
1
IAS
Anthidium manicatum
1
IAS
Coelioxys porterae
1
IAS
Coelioxys sayi
1
IAS
Hylaeus fedorica
1
IAS
Hylaeus leptocephalus
1
IAS
Lasioglossum coreopsis
1
IAS
Lasioglossum cressonii
1
IAS
Lasioglossum ephialtum
1
IAS
Lasioglossum oblongum
1
IAS
Lasioglossum oenotherae
1
IAS
Lasioglossum paradmirandum
1
IAS
Lasioglossum pectorale
1
IAS
Lasioglossum smilacinae
1
IAS
Lasioglossum subviridatum
1
IAS
Megachile brevis
1
IAS
Megachile georgica
1
IAS
Megachile pugnata
1
IAS
Melissodes agilis
1
IAS
Melissodes denticulatus
1
IAS
Nomada lehighensis
1
IAS
Ptilothrix bombiformis
1
IAS
Sphecodes atlantis
1
Lord Stirling
Bombus impatiens
340
Lord Stirling
Ceratina calcarata
253
Lord Stirling
Halictus ligatus
215
Lord Stirling
Bombus griseocollis
201
Lord Stirling
Bombus bimaculatus
152
Lord Stirling
Lasioglossum versatum
125
Lord Stirling
Hylaeus affinis_modestus
81
Lord Stirling
Hylaeus mesillae
74
Lord Stirling
Lasioglossum imitatum
74
Lord Stirling
Augochlorella aurata
65
Lord Stirling
Ceratina strenua
54
Lord Stirling
Lasioglossum paradmirandum
44
Lord Stirling
Lasioglossum hitchensi_weemsi
41
Lord Stirling
Melissodes denticulatus
36
Lord Stirling
Augochlora pura
34
Lord Stirling
Xylocopa virginica
34
Lord Stirling
Agapostemon virescens
21
Lord Stirling
Ceratina dupla
21
Lord Stirling
Ceratina mikmaqi
18
Lord Stirling
Halictus confusus
17
Lord Stirling
Osmia pumila
12
Lord Stirling
Bombus perplexus
8
Lord Stirling
Lasioglossum gotham
8
Lord Stirling
Megachile sculpturalis
8
Lord Stirling
Lasioglossum anomalum
7
Lord Stirling
Megachile frugalis
7
Lord Stirling
Augochloropsis metallica
6
Lord Stirling
Bombus vagans
6
Lord Stirling
Andrena wilkella
5
Lord Stirling
Halictus rubicundus
5
Lord Stirling
Lasioglossum obscurum
5
Lord Stirling
Megachile mendica
5
Lord Stirling
Anthidium oblongatum
4
Lord Stirling
Lasioglossum trigeminum
4
Lord Stirling
Osmia atriventris
4
Lord Stirling
Coelioxys sayi
3
Lord Stirling
Heriades carinata
3
Lord Stirling
Lasioglossum tegulare
3
Lord Stirling
Megachile montivaga
3
Lord Stirling
Agapostemon sericeus
2
Lord Stirling
Lasioglossum atwoodi
2
Lord Stirling
Lasioglossum cressonii
2
Lord Stirling
Lasioglossum oblongum
2
Lord Stirling
Megachile campanulae
2
Lord Stirling
Megachile exilis
2
Lord Stirling
Melissodes subillatus
2
Lord Stirling
Nomada pygmaea
2
Lord Stirling
Anthidium manicatum
1
Lord Stirling
Anthophora terminalis
1
Lord Stirling
Coelioxys modestus
1
Lord Stirling
Hoplitis pilosifrons
1
Lord Stirling
Lasioglossum coeruleum
1
Lord Stirling
Lasioglossum coriaceum
1
Lord Stirling
Lasioglossum illinoense
1
Lord Stirling
Lasioglossum nigroviride
1
Lord Stirling
Lasioglossum oceanicum
1
Lord Stirling
Lasioglossum platyparium
1
Lord Stirling
Lithurgus chrysurus
1
Lord Stirling
Megachile rotundata
1
Lord Stirling
Melissodes desponsus
1
Lord Stirling
Osmia albiventris
1
Lord Stirling
Osmia distincta
1
Lord Stirling
sand wasp_sp
1
Lord Stirling
Stelis louisae
1
MU_2013
Bombus impatiens
151
MU_2013
Ceratina calcarata
34
MU_2013
Halictus ligatus
26
MU_2013
Ceratina strenua
22
MU_2013
Andrena fragilis
20
MU_2013
Hylaeus affinis_modestus
14
MU_2013
Augochlora pura
13
MU_2013
Bombus bimaculatus
10
MU_2013
Lasioglossum oceanicum
10
MU_2013
Xylocopa virginica
10
MU_2013
Lasioglossum gotham
6
MU_2013
Bombus griseocollis
5
MU_2013
Andrena cressonii
4
MU_2013
Megachile mendica
4
MU_2013
Augochlorella aurata
3
MU_2013
Andrena vicina
2
MU_2013
Bombus fervidus
2
MU_2013
Halictus rubicundus
2
MU_2013
Hylaeus mesillae
2
MU_2013
Agapostemon virescens
1
MU_2013
Andrena alleghaniensis
1
MU_2013
Andrena nasonii
1
MU_2013
Andrena nivalis
1
MU_2013
Andrena nuda
1
MU_2013
Andrena platyparia
1
MU_2013
Andrena robertsonii
1
MU_2013
Anthidium manicatum
1
MU_2013
Augochloropsis metallica
1
MU_2013
Ceratina dupla
1
MU_2013
Ceratina mikmaqi
1
MU_2013
Hylaeus modestus
1
MU_2013
Lasioglossum hitchensi_weemsi
1
MU_2013
Lasioglossum imitatum
1
MU_2013
Lasioglossum obscurum
1
MU_2013
Lasioglossum trigeminum
1
MU_2013
Lasioglossum versatum
1
MU_2013
Lasioglossum viridatum
1
RO_2014
Halictus ligatus
74
RO_2014
Bombus impatiens
65
RO_2014
Augochlorella aurata
17
RO_2014
Lithurgus chrysurus
17
RO_2014
Augochlora pura
10
RO_2014
Lasioglossum versatum
10
RO_2014
Augochlorella persimilis
8
RO_2014
Bombus vagans
5
RO_2014
Ceratina calcarata
5
RO_2014
Bombus griseocollis
4
RO_2014
Agapostemon virescens
1
RO_2014
Bombus bimaculatus
1
RO_2014
Bombus fervidus
1
RO_2014
Bombus perplexus
1
RO_2014
Ceratina strenua
1
RO_2014
Halictus confusus
1
RO_2014
Megachile brevis
1
RO_2014
Megachile mendica
1
RO_2014
Megachile rotundata
1
RO_2014
Melissodes agilis
1
RO_2014
Osmia georgica
1
RO_2014
Xylocopa virginica
1
URWA
Ceratina calcarata
256
URWA
Hylaeus affinis_modestus
240
URWA
Hylaeus mesillae
149
URWA
Lasioglossum versatum
127
URWA
Halictus ligatus
125
URWA
Bombus griseocollis
102
URWA
Bombus impatiens
98
URWA
Bombus bimaculatus
62
URWA
Augochlorella aurata
36
URWA
Andrena wilkella
32
URWA
Lasioglossum gotham
32
URWA
Ceratina strenua
29
URWA
Ceratina dupla
26
URWA
Ceratina mikmaqi
25
URWA
Anthidium oblongatum
16
URWA
Augochlora pura
16
URWA
Halictus confusus
14
URWA
Lasioglossum hitchensi_weemsi
14
URWA
Andrena cressonii
13
URWA
Megachile sculpturalis
9
URWA
Bombus perplexus
8
URWA
Osmia pumila
6
URWA
Andrena imitatrix
4
URWA
Andrena robertsonii
4
URWA
Halictus rubicundus
4
URWA
Lasioglossum atwoodi
4
URWA
Megachile gemula
4
URWA
Megachile mendica
4
URWA
Andrena nasonii
3
URWA
Andrena vicina
3
URWA
Anthophora abrupta
3
URWA
Lasioglossum viridatum
3
URWA
Megachile brevis
3
URWA
Megachile frugalis
3
URWA
Osmia distincta
3
URWA
Xylocopa virginica
3
URWA
Andrena pruni
2
URWA
Augochloropsis metallica
2
URWA
Bombus vagans
2
URWA
Hoplitis pilosifrons
2
URWA
Lasioglossum admirandum
2
URWA
Lasioglossum foxii
2
URWA
Lasioglossum tegulare
2
URWA
Osmia bucephala
2
URWA
Agapostemon sericeus
1
URWA
Agapostemon virescens
1
URWA
Andrena hippotes
1
URWA
Andrena nuda
1
URWA
Andrena wilmattae
1
URWA
Anthidiellum notatum
1
URWA
Anthophora terminalis
1
URWA
Heriades carinata
1
URWA
Hoplitis producta
1
URWA
Hoplitis spoliata
1
URWA
Hylaeus leptocephalus
1
URWA
Lasioglossum birkmanni
1
URWA
Lasioglossum callidum
1
URWA
Lasioglossum oblongum
1
URWA
Lasioglossum obscurum
1
URWA
Lasioglossum oceanicum
1
URWA
Lasioglossum smilacinae
1
URWA
Lasioglossum trigeminum
1
URWA
Lithurgus chrysurus
1
URWA
Megachile melanophaea
1
URWA
Megachile montivaga
1
URWA
Megachile rotundata
1
URWA
Melissodes subillatus
1
URWA
Nomada articulata
1
URWA
Nomada bidentate_group
1
URWA
Osmia albiventris
1
URWA
Osmia georgica
1
... (Chen et al., 2020;Hędrzak et al., 2021;Kleijn et al., 2015;Kremen et al., 2002;MacLeod et al., 2020;Simpson et al., 2022;Soliveres et al., 2016;Staton et al., 2022;Sutter et al., 2017;Winfree et al., 2015Winfree et al., , 2018Zhang et al., 2022). For example, MacLeod et al. (2020) studied the overlap in identity and flower preferences between regionally rare species and dominant pollinators in United States (followingKleijn et al.'s (2015) Studies identified in the systematic review that provided quantitative measures of the contribution of rare or endangered species (RES) to agricultural production. ...
Article
Full-text available
Biodiversity underpins ecosystem functions that provide benefits to people, yet the role of rare and endangered species (RES) in supporting ecosystem services is unclear. Thus, it remains controversial whether arguments for conservation that focus on ecosystem services align with the protection of RES. We designed a systematic review protocol to critically assess the evidence for quantitative contributions of RES to terrestrial agricultural production, which is a key driver of biodiversity change and, simultaneously, could suffer from the loss of ecosystem services provided by biodiversity. Our review search criteria required that studies: 1) provide information on RES, 2) focus on an ecosystem service relevant for agriculture; and 3) include a quantitative measure of agricultural production. Surprisingly, we found only four studies that fulfilled these criteria, which was insufficient to perform a meta-analysis of results. Thus, we highlight here the gap in quantitative research, discuss the implications of this knowledge gap for the conservation of RES, and suggest future research directions. We conclude that further quantitative research is urgently needed to better inform conservation and agricultural policies, including research that focuses specifically on RES, incorporates more ecosystem services, and covers a wider range of climatic and socioeconomic contexts.
... Mutualism is widespread in nature and can play a key role in determining individual fitness, species interaction and ecosystem function [1][2][3][4]. In a mutualistic interaction, a species offers commodities to its partner, in exchange for commodities that are impossible to produce [5,6]. ...
Article
Mutualistic interactions between species play a key role in ecosystem function. Theoretical models and laboratory experiments suggest that mutualistic systems can display chaotic population dynamics. Evidence of mutualistic chaos in nature, however, is rare and poorly understood. Here, we show the existence of chaotic behaviour in a mutualistic system. We combine a mathematical model with field data to analyse population fluctuations and the stability of a host-mutualist-exploiter system. Our goal is to understand whether the system can evolve to a steady state and under what environmental conditions. We establish a discrete, seasonally explicit host-mutualist-exploiter model and estimate the parameters using data collected in the field. Model simulations show that the host decreases the ratio of extra reward to host sanction to maintain its higher population density when external environment variation (such as variation in seasonal temperature) is small. The model parameterized with field data predicts that host regulation and environmental variation together, but not individually, can lead to population chaos. Our work provides a new mechanism of population chaos in a mutualistic system and may inspire more interest in exploring chaotic dynamics in natural ecological systems.
Article
Full-text available
Positive biodiversity‐ecosystem functioning (BEF) relationships are predicted to increase in strength when high environmental variability allows for complementarity between resource use strategies in diverse communities. This environmental variability can be represented by spatial or temporal variation in nutrient ratios, but resource use efficiency (RUE) and therefore biomass build up of primary producers might be restricted when nutrient ratios are highly imbalanced (i.e., limitation by one nutrient and beyond optimal ratios for growth). Whereas the linkages between ecosystem functioning, diversity and nutrient availability are theoretically well understood, we lack experimental evidence on how phytoplankton diversity affects resource use and biomass under variable nutrient ratios (N:P ratios). Combining a mesocosm and a microcosm experiment we tested diversity effects on ecosystem functioning by exposing a species diversity gradient generated by the loss of rare species in a natural community to different N:P ratios (uniform vs a gradient). The N:P supply ratio gradient also allowed us to evaluate responses across balanced and imbalanced ratios. We found that increased species diversity led to increased community RUE when supplied a gradient of N:P ratios; but restricted to the highest diversity level. However, diversity did not affect RUE under uniform nutrient ratios. The overall phytoplankton biomass and carbon:nutrient ratios responses to diversity reflected the patterns detected for RUE. Contrary to theoretical predictions, RUE was maintained under highest N:P supply ratios (extreme phosphorous limitation) suggesting that imbalanced N:P ratios do not necessarily decrease function. Thus, we showed that the nutrient context influences diversity effects on RUE and biomass. Synthesis. Overall, our results suggest that the effect of rare phytoplankton species losses on community RUE and biomass can be compensated by the persistent species when nutrient ratios are uniform, but leads to decreases in ecosystem functioning under variable nutrient ratios. This work provides a first attempt for testing interactions between the nutrient context (including concentrations and ratios) and the diversity of (natural) communities experimentally, which is conceptually understood but poorly tested for phytoplankton.
Article
Full-text available
Identifying species that are both geographically restricted and functionally distinct, i.e. supporting rare traits and functions, is of prime importance given their risk of extinction and their potential contribution to ecosystem functioning. We use global species distributions and functional traits for birds and mammals to identify the ecologically rare species, understand their characteristics, and identify hotspots. We find that ecologically rare species are disproportionately represented in IUCN threatened categories, insufficiently covered by protected areas, and for some of them sensitive to current and future threats. While they are more abundant overall in countries with a low human development index, some countries with high human development index are also hotspots of ecological rarity, suggesting transboundary responsibility for their conservation. Altogether, these results state that more conservation emphasis should be given to ecological rarity given future environmental conditions and the need to sustain multiple ecosystem processes in the long-term.
Article
Full-text available
A rich body of knowledge links biodiversity to ecosystem functioning (BEF), but it is primarily focused on small scales. We review the current theory and identify six expectations for scale dependence in the BEF relationship: (1) a nonlinear change in the slope of the BEF relationship with spatial scale; (2) a scale‐dependent relationship between ecosystem stability and spatial extent; (3) coexistence within and among sites will result in a positive BEF relationship at larger scales; (4) temporal autocorrelation in environmental variability affects species turnover and thus the change in BEF slope with scale; (5) connectivity in metacommunities generates nonlinear BEF and stability relationships by affecting population synchrony at local and regional scales; (6) spatial scaling in food web structure and diversity will generate scale dependence in ecosystem functioning. We suggest directions for synthesis that combine approaches in metaecosystem and metacommunity ecology and integrate cross‐scale feedbacks. Tests of this theory may combine remote sensing with a generation of networked experiments that assess effects at multiple scales. We also show how anthropogenic land cover change may alter the scaling of the BEF relationship. New research on the role of scale in BEF will guide policy linking the goals of managing biodiversity and ecosystems. We address the challenge of scale for biodiversity and ecosystem functioning (BEF) research. We review current theory and identify six expectations for scale dependence in the BEF relationship. We suggest directions for synthesis that combine theoretical and empirical methods and suggest their application to human transformed landscapes.
Article
Full-text available
Stimulated cells and cancer cells have widespread shortening of mRNA 3’-untranslated regions (3’UTRs) and switches to shorter mRNA isoforms due to usage of more proximal polyadenylation signals (PASs) in introns and last exons. U1 snRNP (U1), vertebrates’ most abundant non-coding (spliceosomal) small nuclear RNA, silences proximal PASs and its inhibition with antisense morpholino oligonucleotides (U1 AMO) triggers widespread premature transcription termination and mRNA shortening. Here we show that low U1 AMO doses increase cancer cells’ migration and invasion in vitro by up to 500%, whereas U1 over-expression has the opposite effect. In addition to 3’UTR length, numerous transcriptome changes that could contribute to this phenotype are observed, including alternative splicing, and mRNA expression levels of proto-oncogenes and tumor suppressors. These findings reveal an unexpected role for U1 homeostasis (available U1 relative to transcription) in oncogenic and activated cell states, and suggest U1 as a potential target for their modulation. U1 snRNP is a key regulator of mRNA biogenesis through its roles in splicing, and transcription and 3’-end processing. Here the authors show a tumor suppressor-like function of U1 snRNP using in vitro cell migration/invasion assays and transcriptome profiling.
Article
Full-text available
The ecological importance of common species for many ecosystem processes and functions is unquestionably due to their high abundance. Yet, the importance of rare species is much less understood. Here we take a theoretical approach, exposing dynamical models of ecological networks to small perturbations, to explore the dynamical importance of rare and common species. We find that both species types contribute to the recovery of communities following generic perturbations (i.e. perturbations affecting all species). Yet, when perturbations are selective (i.e. affects only one species), perturbations to rare species have the most pronounced effect on community stability. We show that this is due to the strong indirect effects induced by perturbations to rare species. Because indirect effects typically set in at longer timescales, our results indicate that the importance of rare species may be easily overlooked and thus underrated. Hence, our study provides a potential ecological motive for the management and protection of rare species.
Chapter
Full-text available
Concern about the functional consequences of unprecedented loss in biodiversity has prompted biodiversity-ecosystem functioning (BEF) research to become one of the most active fields of ecological research in the past 25 years. Hundreds of experiments have manipulated biodiversity as an independent variable and found compelling support that the functioning of ecosystems increases with the diversity of their ecological communities. This research has also identified some of the mechanisms underlying BEF relationships, some context-dependencies of the strength of relationships, as well as implications for various ecosystem services that mankind depends upon. In this paper, we argue that a multitrophic perspective of biotic interactions in random and non-random biodiversity change scenarios is key to advance future BEF research and to address some of its most important remaining challenges. We discuss that the study and the quantification of multitrophic interactions in space and time facilitates scaling up from small-scale biodiversity manipulations and ecosystem function assessments to management-relevant spatial scales across ecosystem boundaries. We specifically consider multitrophic conceptual frameworks to understand and predict the context-dependency of BEF relationships. Moreover, we highlight the importance of the eco-evolutionary underpinnings of multitrophic BEF relationships. We outline that FAIR data (meeting the standards of findability, accessibility, interoperability, and reusability) and reproducible processing will be key to advance this field of research by making it more integrative. Finally, we show how these BEF insights may be implemented for ecosystem management, society, and policy. Given that human well-being critically depends on the multiple services provided by diverse, multitrophic communities, integrating the approaches of evolutionary ecology, community ecology, and ecosystem ecology in future BEF research will be key to refine conservation targets and develop sustainable management strategies.
Article
Full-text available
Background Intraspecific variation in foraging niche can drive food web dynamics and ecosystem processes. In particular, male and female animals can exhibit different, often cascading, impacts on their interaction partners. Despite this, studies of plant-pollinator interaction networks have focused on the partitioning of the floral community between pollinator species, with little attention paid to intraspecific variation in plant preference between male and female bees. We designed a field study to evaluate the strength and prevalence of sexually dimorphic foraging, and particularly resource preferences, in bees. Study design We observed bees visiting flowers in semi-natural meadows in New Jersey, USA. To detect differences in flower use against a shared background of resource (flower) availability, we maximized the number of interactions observed within narrow spatio-temporal windows. To distinguish observed differences in bee use of flower species, which can reflect abundance patterns and sampling effects, from underlying differences in bee preferences, we analyzed our data with both a permutation-based null model and random effects models. Findings We found that the diets of male and female bees of the same species were often dissimilar as the diets of different species of bees. Furthermore, we demonstrate differences in preference between male and female bees. We show that intraspecific differences in preference can be robustly identified among hundreds of unique species-species interactions, without precisely quantifying resource availability, and despite high phenological turnover of both bees and plant bloom. Given the large differences in both flower use and preferences between male and female bees, ecological sex differences should be integrated into studies of bee demography, plant pollination, and coevolutionary relationships between flowers and insects.
Article
Full-text available
Approximately 25 years ago, ecologists became increasingly interested in the question of whether ongoing biodiversity loss matters for the functioning of ecosystems. As such, a new ecological subfield on Biodiversity and Ecosystem Functioning (BEF) was born. This subfield was initially dominated by theoretical studies and by experiments in which biodiversity was manipulated, and responses of ecosystem functions such as biomass production, decomposition rates, carbon sequestration, trophic interactions and pollination were assessed. More recently, an increasing number of studies have investigated BEF relationships in non‐manipulated ecosystems, but reviews synthesizing our knowledge on the importance of real‐world biodiversity are still largely missing. I performed a systematic review in order to assess how biodiversity drives ecosystem functioning in both terrestrial and aquatic, naturally assembled communities, and on how important biodiversity is compared to other factors, including other aspects of community composition and abiotic conditions. The outcomes of 258 published studies, which reported 726 BEF relationships, revealed that in many cases, biodiversity promotes average biomass production and its temporal stability, and pollination success. For decomposition rates and ecosystem multifunctionality, positive effects of biodiversity outnumbered negative effects, but neutral relationships were even more common. Similarly, negative effects of prey biodiversity on pathogen and herbivore damage outnumbered positive effects, but were less common than neutral relationships. Finally, there was no evidence that biodiversity is related to soil carbon storage. Most BEF studies focused on the effects of taxonomic diversity, however, metrics of functional diversity were generally stronger predictors of ecosystem functioning. Furthermore, in most studies, abiotic factors and functional composition (e.g. the presence of a certain functional group) were stronger drivers of ecosystem functioning than biodiversity per se. While experiments suggest that positive biodiversity effects become stronger at larger spatial scales, in naturally assembled communities this idea is too poorly studied to draw general conclusions. In summary, a high biodiversity in naturally assembled communities positively drives various ecosystem functions. At the same time, the strength and direction of these effects vary highly among studies, and factors other than biodiversity can be even more important in driving ecosystem functioning. Thus, to promote those ecosystem functions that underpin human well‐being, conservation should not only promote biodiversity per se, but also the abiotic conditions favouring species with suitable trait combinations.
Article
Full-text available
Human-induced biodiversity change impairs ecosystem functions crucial to human well-being. However, the consequences of this change for ecosystem multifunctionality are poorly understood beyond effects of plant species loss, particularly in regions with high biodiversity across trophic levels. Here we adopt a multitrophic perspective to analyze how biodiversity affects multifunctionality in biodiverse subtropical forests. We consider 22 independent measurements of nine ecosystem functions central to energy and nutrient flow across trophic levels. We find that individual functions and multifunctionality are more strongly affected by the diversity of heterotrophs promoting decomposition and nutrient cycling, and by plant functional-trait diversity and composition, than by tree species richness. Moreover, cascading effects of higher trophic-level diversity on functions originating from lower trophic-level processes highlight that multitrophic biodiversity is key to understanding drivers of multi-functionality. A broader perspective on biodiversity-multifunctionality relationships is crucial for sustainable ecosystem management in light of non-random species loss and intensified biotic disturbances under future environmental change.
Article
Conservation aims to preserve species and ecosystem services. If rare species contribute little to ecosystem services, yet are those most in need of preservation, tradeoffs may exist for these contrasting objectives. However, little attention has focused on identifying how, when, and where rare species contribute to ecosystem services and at what scales. Here, we review distinct ways that ecosystem services can positively depend on the presence, abundance, disproportionate contribution or, counterintuitively, the scarcity of rare species. By contrast, ecosystem services are less likely to depend on rare species that do not have a unique role in any service or become abundant enough to contribute substantially. We propose a research agenda to identify when rare species may contribute significantly to services.