Technical ReportPDF Available

Neonicotinoids in California Their Use and Threats to the State's Aquatic Ecosystems and Pollinators, with a Focus on Neonic-Treated Seeds


Abstract and Figures

Neonicotinoids in California
Their Use and Threats to the States Aquatic Ecosystems
and Pollinators, with a Focus on Neonic-Treated Seeds
Pierre Mineau; Pierre Mineau Consulting
Pierre Mineau, PhD
Principal Senior Scientist at Pierre Mineau Consulting
Adjunct Professor, Department of Biology, Carleton University
Subject Editor, Earth Systems and Environmental Sciences, Elsevier Publishing
Contact:; +1 (613) 853-2013
Professional website:
Google scholar link:
Access to open source publications:le/Pierre_Mineau/contributions?ev=prf_act
Executive summary ..........................................................................................................................................................................1
Acknowledgments ...........................................................................................................................................................................2
1. Introduction .............................................................................................................................................................................2
2. What is the potential total nitroguanidine neonic use in California, and what proportion comes from seed treatments? ......... 3
2.1. e role of neonic-treated seeds in agriculture ......................................................................................................................... 3
2.2. Comparing the total potential neonic-treated seed use to PUR estimates of tracked neonic uses .................................... 4
2.2.1. Calculating the total potential neonic-treated seed use in California ........................................................................ 5
2.2.2. Comparing the total potential neonic-treated seed use to PUR data for non-seed neonic uses .......................... 15
2.3. Estimating the actual total neonic-treated seed use in California ........................................................................................ 15
3. Aquatic contamination from nitroguanidine neonicotinoids in California .............................................................................16
3.1. At what level do neonics cause ecological harm? .................................................................................................................... 16
3.2. Predicted contamination from seed treatments relative to other application methods ..................................................... 17
3.3. Evidence of neonic contamination in California surface water data and likely environmental impacts ........................ 20
3.3.1. Continued monitoring of imidacloprid at sites from the Starner and Goh (2012) study ...................................... 23
3.3.2. Clothianidin and thiamethoxam sampling .................................................................................................................. 24
4. Pollinator impacts from nitroguanidine neonicotinoid seed treatments in California ...............................................................26
4.1 Recent regulatory reviews of neonics and pollinators ............................................................................................................ 26
4.2. What does the CaDPR assessment conclude with respect to neonic-treated seeds? .......................................................... 26
4.3. What does the USEPA assessment conclude with respect to neonic-treated seeds? .......................................................... 27
4.4. Deciencies in USEPAs assessment of pollinator risk from neonic-treated seeds ............................................................. 28
4.4.1. e USEPA assessment fails to adequately consider risks to pollinators beyond honey bees ............................... 28
4.4.2. e USEPA assessment underestimates nectar and pollen contamination that results
from the use of seed treatments ..................................................................................................................................... 31
4.4.3. e USEPA assessment ignores critical exposure routes ............................................................................................ 33
4.5. Risks and impacts to pollinators from neonic-treated seeds are not restricted to the crop area ...................................... 37
4.6. Harmful impacts from neonic-treated seeds have already been demonstrated .................................................................. 37
4.7. Conclusions of the pollinator assessment ................................................................................................................................ 38
Literature cited ..............................................................................................................................................................................39
Appendix 1. Data standardization procedures to build an estimate of potential neonic
seed treatment uses in California from state and USDA sources .....................................................................................................46
Appendix 2. Tabulation of how oen each crop or crop grouping appeared as an approved neonic seed
treatment for potential use in California from 106 separate pesticide labels downloaded in March 2020 ....................................... 47
Appendix 3. County by county comparison of potential seed treatment use of nitroguanidine
neonics versus their reported use according to the PUR system ..................................................................................................... 49
Appendix 4. What we know about the current use of neonic seed treatments in California ............................................................ 51
Appendix 5. County by county comparison of total cropland considered in this report
compared to the potential area with neonic seed treatments. .........................................................................................................53
Appendix 6. Summary of pollen and nectar neonic residue levels following seed treatment
use in registrant studies reviewed in the USEPA assessment ..........................................................................................................55
Neonicotinoid insecticides (neonics) are the most widely used insecticides in the United States—and the largest proportion
of that use comes from neonic-coated crop seeds. The U.S. Environmental Protection Agency (USEPA) and the California
Department of Pesticide Regulation (CaDPR) do not regulate treated seeds as “pesticides, allowing them to escape scrutiny
and tracking requirements applicable to other pesticide uses. Increasingly, farmers, and even agronomists, are unaware of the
extent to which seed treatment insecticides are being used. The lack of accurate information limits the ability to assess neonic-
treated seeds’ full impacts on agricultural production and economics, pest resistance issues, and the environment.
This report explores the use and impacts of neonics in California with a focus on the possible consequences of ignoring seed
treatment applications, particularly with regards to impacts on aquatic life and pollinators. The report finds that neonic-
treated seeds have the potential to be the single largest use of neonics in California, making them, therefore, also one of the
largest insecticide uses. It also finds that the possible detrimental impacts of neonic-treated seeds on California’s aquatic
ecosystems and pollinators are wholly uncharacterized and greatly under-appreciated.
Neonic-treated seed use is not tracked in California’s Pesticide Use Reporting database (PUR). As a result, neonic-treated seed
use has never been estimated in California as opposed to all other states where the data was tracked by the U.S. Geological
Survey until the agency discontinued tracking seed treatments in 2014. Accordingly, to examine the possible impact of neonic-
treated seeds on California’s environment—both now and in the future—this report estimates the total potential use of
neonics on crop seeds in the State based on 2016-2017 crop data. In California, exact use patterns are uncertain, but national
trends suggest use of neonic-treated seeds is widespread in field crops (such as corn, cotton, and wheat), and likely on the rise
in others. The report concludes that:
If seed treatments were fully used on crops where they are allowed, the amount of neonics applied as seed treatments
would equal 512,000 pounds annually. This total exceeds the 410,000 pounds of neonics that are applied by other means
and reported through the PUR.
This potential use of neonic-treated seeds would cover roughly 76% of the total cropland area in California,
approximately 4 million acres.
This potential neonic-treated seed use has implications for water quality and ecosystem health. In its modeling approach,
USEPA greatly underestimates aquatic contamination from neonic-treated seeds; ample independent field evidence already
shows that use of neonic-treated seeds results in neonic levels in water sufficient to cause injury to aquatic habitats. If the use
of neonic-treated seeds in California were even a fraction of the total potential use estimated in this report, contamination of
and harm to aquatic habitats would be expected.
Regardless of the source, neonics frequently appear in California water at levels where we would expect to see damaging
ecosystem-wide impacts. These impacts include widespread losses of aquatic insects and crustacea accompanied by knock-on
effects on consumer species such as fish, amphibians, birds, and some mammals such as bats. At agricultural sites monitored
by CaDPR, concentrations of the neonic imidacloprid, where detected, always exceeded the USEPA concentration benchmark
where ecological damage is expected. At many sites, the benchmark was exceeded by 10X or even 100X. At sites described
in this report—for example, tributaries to the Salinas River in Monterey County—every single sample taken over an 8-year
period detected imidacloprid 10X above the USEPA ecological damage threshold. This data alone indicate a very high
probability that neonics are causing ecosystem-wide damage.
Clothianidin and thiamethoxam are neonic active ingredients commonly used as seed treatments, and water sampling for
these chemicals in California has not been adequate. CaDPR does not include these chemicals within a comprehensive
agricultural sampling program or track seed treatment products more generally, creating a serious data gap that deprives
scientists and policy makers of key information about the likely significant and damaging impacts of neonic-treated seed use.
Indeed, even despite minimal sampling, evidence suggests that clothianidin and thiamethoxam residues are likely higher than
imidacloprid residues at many agricultural sites. Given that most aquatic systems within agricultural regions will be exposed
to the three nitoguanidine products in combination, the combined concentrations of all neonic active ingredients must be
considered in assessing injury to the aquatic environment.
USEPA and CaDPR fail to properly characterize the risk of neonic seed treatments to bees, especially wild bees. In particular,
the final USEPA (2020) and CaDPR (2018) pollinator risk assessments: (1) underestimate risks to wild bee species and other
pollinators by relying on honey bee colony health and survival as a proxy for pollinator health generally; (2) underestimate
nectar and pollen contamination levels following the use of neonic-treated seeds by assuming that the majority of crop
species will have residue values at the low end of the measured spectrum; (3) ignore risks of dust from neonic-treated seeds
at planting, despite ample evidence that this route of exposure is highly relevant; (4) ignore exposures of bees and other
pollinators to neonic-contaminated water—including guttation fluid and puddles in or near fields sown with neonic-treated
seeds—despite existing field estimates that show that these routes of exposure can completely dwarf the routes that have been
formally assessed; (5) ignore risks from neonic uses on crops deemed unattractive to honey bees, despite evidence that neonic
residues migrate into adjoining areas, including wildflowers bordering neonic-seeded fields that can contain neonic levels
exceeding those in the field proper; (6) exclude available peer-reviewed literature in favor of industry studies; and (7) ignore
the growing amount of field data which now links the use of neonic-treated seeds to pollinator failure on a landscape scale.
The USEPA and CaDPR therefore fail to account for the likely considerable and damaging effect that neonic-treated seeds are
having on California’s pollinator populations.
All of these impacts on aquatic systems and pollinators have to be weighed against what appear to be questionable benefits of
many seed treatment products to California agriculture.
The author would like to thank Ann Balke, Daniel Raichel, and Jennifer Sass from the Natural Resources Defense Council
(NRDC) for their contributions and input into this report and to NRDC for providing the financial support to produce it.
A thank you also goes to those California crop specialists who responded to the author’s queries. The views expressed in this
report remain those of the author.
There has been a large increase, over the last few decades, in the use of insecticidal seed treatments. This has happened in
parallel with the introduction of neonicotinoid insecticides or “neonics,” which have marked systemic activity. The U.S.
Environmental Protection Agency (USEPA), California Department of Pesticide Registration (CaDPR), and other regulatory
bodies such as the Canadian Pest Management Regulatory Agency (PMRA) do not directly regulate pesticide-treated seeds
under their respective laws. This regulatory loophole or grey area has been of increasing concern to several researchers
(Douglas et al. 2015, Hitaj et al. 2020), has already been the subject of articles in popular media (e.g., Allington 2020), and
has resulted in a formal citizen’s petition from a coalition of farmers, beekeepers, and not-for-profit organizations to USEPA
in 2017 (Jenkins 2017). Yet despite their dominance of insecticide markets, neonic seed treatments are, for the most part,
untracked. Agronomists are generally unaware to what extent insecticide seed treatments are being used in the crops they
work on, and even farmers themselves are often in the dark.
The main objective of this report is to consider the use of neonics, especially the three main nitroguanidine active ingredients:
clothianidin, imidacloprid, and thiamethoxam, which represent the bulk of seed treatment uses. For at least two of these active
ingredients (clothianidin and thiamethoxam), seed treatment use is the principal way in which they are released into the
environment—at least nationally.
Because of its robust pesticide evaluation and registration system as well as arguably the best pesticide tracking system in
the world (NRC 2015), the State of California occupies an enviable position in the realm of pesticide regulation and data
collection. Yet, it appears that, even in their case, seed treatments are a major blind spot, as the state does not register or track
the use of pesticide treated seeds.
Given California’s diverse and rich agriculture and its relatively robust data on pesticide use and impacts, the state provides
a good test case for understanding how widely neonic seed treatments may be used and what the impact of those treatments
and other neonic uses have on water quality and natural systems, with a particular focus on pollinating insects. Accordingly,
this report proposes to: (1) construct a snapshot of California agriculture to estimate the amount of total neonic use
seed treatments could represent (Section 2); (2) review what is known about the current degree of neonic surface water
contamination (Section 3) and; (3) review the possible consequences of these undocumented sources of neonics to pollinator
species (Section 4).
2.1. The role of neonic-treated seeds in agriculture
The use of insecticidal seed treatments has dramatically increased in the last few decades, in parallel with the introduction of
neonicotinoid insecticides (Hitaj et al. 2020). Neonics, especially those in the nitroguanidine group (including imidacloprid,
clothianidin, and thiamethoxam),1 are highly water soluble, systemic, persistent, and broadly toxic to a wide range of
terrestrial and aquatic invertebrates. It has been estimated that, by 2008, neonics already constituted 80% of the insecticide
seed treatment market (Jeschke et al. 2011), which itself accounts for an increasing proportion of all insecticides used in
agriculture. Therefore, nationally at least, a large proportion of current insecticide use is in the form of neonic seed treatments.
As state governments increasingly restrict or prohibit the remaining uses of organophosphorus insecticides (e.g., chlorpyrifos),
the use of insecticidal seed treatments is likely to increase further. This shift has been criticized as pesticide seed treatments
are, by design, prophylactic—meaning that treatments on an unknown portion of seeds serve no purpose whatsoever. Indeed,
research indicates that the benefits of treatment for some crops are slim to non-existent in many cases (MacFadyen et al. 2014,
Budge et al. 2015, Douglas and Tooker 2015, 2016, Douglas et al. 2015, Krupke et al. 2017, Alford and Krupke 2017, Hokkanen
et al. 2017, Labrie et al. 2020), or can easily be replaced by existing cultural or other methods (Veres et al. 2020). Agronomists
and agricultural researchers increasingly recommend against the prophylactic use of the insecticides, especially in major field
crops, but the practice persists, in part because pesticide manufacturers also control the seed market.
Despite their dominance of insecticide markets, neonic seed treatments are, for the most part, untracked and ignored.2
Through 2014, the private polling company Kynetek was the only entity collecting information on seed treatment use in
the United States. The company, however, stopped in 2015, in part because of the increasing difficulty of collecting the
information (Hitaj et al. 2020). In particular, consolidation of the seed and pesticide manufacturers has led to the bundling
of pesticides with seed stock, so that growers are increasingly unaware that they are using a neonic on their field (and
therefore unable to accurately answer a survey) (Hitaj et al. 2020). Since 2015, the U.S. Department of Agriculture (USDA)
has attempted to fill the information gap through their annual ARMS survey (Agricultural Resource Management Survey),
but the effort has been fraught with methodological problems and the survey has been steadily losing effectiveness (Hitaj et
al. 2020). Further, USEPA and state regulatory bodies exempt treated seeds from regulation as “pesticides” and do not require
the collection of use data. Accordingly, the agencies responsible for the regulation of neonics—the most popular insecticides
in the U.S.—fundamentally lack data regarding their primary use (USEPA 2020a, b). As we move further away in time from
the last 2014 Kynetek survey, our understanding of the actual neonic use on crop seeds—and, thereby, actual insecticide use in
agriculture more generally—becomes less and less clear.
The U.S. Geological Survey (USGS), through the National Water-Quality Assessment’s (NAWQA) “Pesticide National
Synthesis Project,” purchases annual pesticide use surveys from the Kynetec company and makes a summarized version
available to the public. These summaries show a sharp drop in estimated neonic use from 2014 (the last year Kynetec included
seed treatments in its surveys) to 2015. This drop has been used to infer what proportion of total neonic use was the result
of seed treatment use. In previous reports (Mineau 2019a, b), it was estimated that seed treatment use accounted for 58% of
imidacloprid, 75% of thiamethoxam, and 92% of clothianidin nationwide. Hitaj et al. (2020) also proposed this one-time
method to estimate seed treatment uses.
However, it is now apparent that these estimates are underestimates because pre-2015 USGS data for California—an
important contributor to U.S. national agricultural production—did not include seed treatment uses. California is the only
state where USGS used a source other than commercial polling—specifically, the CaDPR’s Pesticide Use Reporting (PUR)
database, which does not track pesticide seed treatment use.
California is unique among U.S. states, indeed worldwide, in that it requires pesticide users (aside from consumers) to log
any pesticide application in the PUR system—providing area treated, crop, product, and location. This has typically led
1 Dinotefuran is also a nitroguanidine neonic but is not registered for seed treatment use.
2 Although seed treatment products (i.e., the products used to treat the seeds) must still undergo registration.
agronomists and other researchers to believe that that PUR data provides a “good handle” on pesticide use in the state. The
PUR data have been key to investigating relationships between various pesticides and health outcomes (NRC 2015). However,
since the PUR does not track seed treatment use, this confidence might be misplaced.
California has a complex and diverse agricultural environment with a large number of possible uses for seed treatments. One
objective of this report is to assess how much the absence of seed treatment uses from the PUR system might underestimate
the use of neonic insecticides—in particular, imidacloprid, clothianidin, and thiamethoxam.3 Given the link between
neonics—and seed treatment products specifically—to honey bee and other pollinator losses, as well as the increasing reports
of broad aquatic contamination from neonic-treated seed use, the question of how much these treated seeds are being used
is an important one. As pointed out by Hitaj et al. (2020), the lack of accurate information on pesticide use limits our ability
to answer questions regarding agricultural production and economics, pest resistance issues, and the full nature of neonics’
environmental risks and harms.
2.2. Comparing the total potential neonic-treated seed use to PUR estimates
of tracked neonic uses
In this section, I estimate the potential amount of nitroguanidine neonics that could be used in California if crops for which
neonic seed treatments were permitted made full use of those treated seeds. Full deployment of seed treatments where allowed
may seem like an extreme case at first blush, but for some major field crops, neonic seed treatment use has become near total,
while others trend in that direction. Over the last decade, there has been an exponential increase in the use of neonics in seed
treatments, and the list of crops for which neonic seed treatment products are not registered has greatly diminished. Estimates
of the extent of seed treatment use, when available, is therefore likely to be rapidly outdated. The extent to which seed
treatment use is covered in the agricultural online press (personal observation) suggests that—despite the agronomic concerns
surrounding prophylactic insecticide use and numerous studies showing neonic-treated field crop seeds fail to produce
monetary benefits for farmers—they are being aggressively marketed, especially as older pesticide chemistries are being retired
(Figure 1).
Figure 1. Excerpt from “The Guide to Seed Treatment Stewardship,” an industry bulletin endorsed by the National
Corn Growers Association, Crop Life America, the American Soybean Association, American Seed Trade Association,
National Cotton Council of America, Agricultural Retailers Association, and National Association of Wheat Growers.
3 Another neonicotinoid chemical, acetamiprid, has a few seed treatment registrations also—primarily canola and mustard. Acetamiprid’s market share is unknown at this
point, but considered to be small (USEPA 2020k).
2.2.1. Calculating the total potential neonic-treated seed use in California Step 1: Building a snapshot of California agriculture
The first step in approximating the total possible annual neonic-treated seed use in California is to estimate the total
planted acreage in California where neonic-treated seed could be used in a given year. To create this “snapshot” of California
agriculture, the year 2016 was chosen as the model year, as, when this project started, it was the most recent year for which
almost all county reports from the various agriculture commissioners were available, and for which USGS4 had completed its
analysis of neonic applications through the PUR system.
Crop information was downloaded from all county reports for the year 2016. The procedures and assumptions made in order
to standardize the data are detailed in Appendix 1.
In many counties, data for various vegetable crops were combined for the purpose of reporting harvested acreages. For many
crops, we therefore turned to the 2017 USDA agricultural census, which provided county-specific acreages for many of the
main vegetable crops. Our final “snapshot” therefore combined the 2016 and 2017 data.
USDA also developed field crop estimates for 2016. Because field crops (sometimes major ones) were frequently combined
in the various state reports, USDA totals were useful to disaggregate the different field crop types. However, USDA data also
combine counties and even agricultural districts together, preventing a clear geographical separation of the data. We therefore
used a combination of state statistics and USDA estimates to arrive at totals for as many counties and individual field crops as
possible. California totals from the various USDA-reported estimates were then used to compute the geographically undefined
part of the field crop. These “residuals” are given in Table 1.
For some field crops, county accounting appears to be reasonable; not so for barley and oats where a large residual exists. In
those cases, acreages were not available from county reports, usually because they were combined with other field crops. Those
residual acreages were included in our estimates of potential seed treatment use but not assigned to any county or agricultural
All of the compiled acreages and USDA data refer to harvested acreages. For the purpose of calculating possible seed treatment
use, seeded acreages are more relevant than harvested acreages. Where crop failure occurs or where the crop is used as a cover
crop (e.g., oats), the planted area can greatly exceed the harvested area. Therefore, USDA field crop totals for California were
used also to work out a ratio of seeded to harvested crops (Table 2). For all crops other than those listed in the table, it was
assumed that the planted acreage was equal to the harvested acreage. This will underestimate plantings and, therefore, seed
treatment use to some extent.
Table 1. Comparison of USDA-reported acreages for 2016 to the summed total acreages reported in 2016 county
agricultural reports for several surveyed field crops.
Residual acreage not included
in county totals (acres)
Proportion this represents of the total acreage
for that crop harvested in California
Barley 24,544 40.9%
Cotton 7,817 3.6%
Corn, grain 1,348 1.3%
Oats 6,563 59.7%
Rice 1,725 0.3%
Sunflower 2,404 5.2%
4 Data obtained from:
Table 2. Ratio of planted to harvested 2016 crop acres for California.a
Area planted
Area harvested
factor Ye a r
Release date
of data
Barley 85,000 60,000 1.417 2016 Mar-18
Cotton 218,000 216,000 1.009 2016 May-18
Dry beans 50,000 49,000 1.020 2016 Apr-18
Corn, total 420,000 415,000 1.012 2016 Feb-18
Oats 110,000 11,000 10.000 2016 Mar-18
Rice 541,000 536,000 1.009 2016 Mar-18
Sunflower 46,600 46,000 1.013 2016 Feb-18
Wheat, total 480,000 217,000 2.212 2016 Mar-18
Safflower 62,000 61,500 1.008 2016 Feb-19
Sugar beet 25,300 25,200 1.004 2016 Feb-19
Potato 39,400 38,300 1.029 2016 Feb-19
Sweet potato 20,000 20,000 1.000 2016 Feb-19
a Data obtained from:;
Crop statistics could not be found for a number of crops for which neonic seed treatments are registered for use in California
(CaDPR 2018). This is in large part because county statistics amalgamated several crops when reporting acreages. These
include: amaranth, arrowroot and leren, arugula, borage, buckwheat, canola and rapeseed, cardoon, cassava, crambe, endive,
edible flowers (e.g., chrysanthemum), fennel, flax, ginger, ginseng, kohlrabi, lupins, pearl and proso millet, all mustards,
purslane, sorrel, soybean, peanuts, teosinte, and turmeric. This will result in an underestimate of potential seed treatment uses. Step 2: Obtaining label rates for registered seed treatments
The second step in approximating neonic-treated seed use is estimating what amount of neonic coating could appear on
a given treated crop seed. To achieve this, all registered seed treatment labels were examined, and, for each registered crop,
maximum rates of application were retained. It is standard practice in regulatory assessments to assume maximum labeled
rates (e.g., CaDPR 2018).5
Depending on seed type and product, label rates vary a great deal as to how application rates are expressed—e.g., ounce of
product per 100 lbs of seed or per a certain number of seeds (typically 80,000, or 100,000) or amount of active ingredient
(usually in milligrams) per individual seed. For our purposes, all rates were converted to either pounds (lbs) of active
ingredient (a.i.)/100 lbs of seed or milligrams (mg) of a.i./seed. Given known seed weights and a given seeding density,
the amounts of a.i. applied per acre can then be calculated for any situation. To arrive at a rate in lbs a.i./100 lbs for liquid
formulations, the specific gravity (in lbs a.i./gal) and the application rates were retained and used as shown in Equation 1.
Equation 1
lbs a.i./100 lbs seed = (lbs a.i./gal) / (128 fl oz/gal) * (fl oz/100 lbs seed)
5 e vast majority of seed treatment products registered by USEPA contain only one neonic active ingredient. A few registered products, however, contain more than one
neonic—e.g., Sepresto 75 WS (EPA reg. no. 264-1081). For the purpose of rate calculations in this report, the most dominant neonic active ingredient on label was chosen.
For solid formulations, the percent proportion of active ingredient and rate in ounces were used as shown in Equation 2.
Equation 2
lbs a.i./100 lbs seed = (% a.i.) * (oz/100 lbs seed)
To arrive at an application rate (in mg) per seed, calculations were performed as per Equation 3.
Equation 3
mg a.i./seed = (fl oz per specified number of seeds) / (specified number of seeds) * (lbs a.i./gal) /
(128 fl oz/gal) * (453592 mg/lb)
Excluding any label with a prohibition of planting the treated seed in California, we documented 106 different extant EPA
labels for seed treatments in various crops—1 acetamiprid label, 12 clothianidin labels, 57 imidacloprid labels and 36
thiamethoxam labels. We also referred to CaDPR (2018), which carried out a similar analysis of California seed treatment uses.
Maximum labeled application rates are provided in Table 3, expressed either as lbs a.i./100 lbs seed or mg a.i./seed for reasons
detailed above. Although there tends to be a convergence of labels towards a given rate, it can sometime vary two-or-three-
fold, presumably in response to different pests for which the product is registered. Higher rates are usually labeled, for
“extended protection” or some such term. It is widely believed that application rates have increased over time (e.g., Douglas
and Tooker 2015) and, therefore, the higher registered rate is likely often the most popular. This is also borne out by the fact
that the higher rates tend to appear more often on labels.
Table 3. Maximum application rates applied to seed and allowed in California.
Crop or crop
Clothianidin Imidacloprid Thiamethoxam
lbs a.i./
100 lbs seed mg/seed
lbs a.i./
100 lbs seed mg/seed
lbs a.i./
100 lbs seed mg/seed
alfalfa 0.001
amaranth 0.050 1.20
arrowroot 0.010
artichoke 0.010
Asian vegetables 0.010 0.125 0.050 1.20
barley 0.070 0.094 0.052
bean 0.010 0.750 0.125 0.050
borage 1.00 0.040
broccoli 0.070 2.12 0.100
Brussel sprout 0.100
buckwheat 0.070
cabbage 2.12 1.00 0.100
canola 0.406 1.00 0.404
cardoon 1.20
carrot 0.010 0.120 0.050
cassava 0.010
celery 1.20
chard 1.20
Table 3. Maximum application rates applied to seed and allowed in California.
Crop or crop
Clothianidin Imidacloprid Thiamethoxam
lbs a.i./
100 lbs seed mg/seed
lbs a.i./
100 lbs seed mg/seed
lbs a.i./
100 lbs seed mg/seed
chicory 0.050 1.20
corn 0.070 1.25 0.528 1.34 0.220 1.25
cotton 0.424 0.502 0.375
crambe 0.400
cucumber 0.750
endive 0.050 1.20
edible flowers 0.010 0.050 1.20
fennel 1.20
flax 1.00 0.400
ginger 0.010
greens 0.406 0.050 1.00 1.20
herbs 0.050 1.20
kale 0.100
kohlrabi 0.100
leek 0.360
lettuce 1.42 1.20
lentil 0.125 0.050
lupin 0.010 0.125 0.050
melon 0.750
millet 0.070 0.250
mustard 1.00 0.400
oats 0.070 0.094 0.051
onion 0.320 0.200
peanuts 0.062 0.045 0.300
peas 0.125 0.050
potato 0.0125 0.0125 0.006
rhubarb 1.20
rice 0.141
rye 0.070 0.094 0.051
safflower 1.00 0.500 0.400
sorghum 0.200 0.250 0.199 0.093
sorrel 0.050 1.20
soybean 0.050 0.130 0.125 0.075 0.151
Table 3. Maximum application rates applied to seed and allowed in California.
Crop or crop
Clothianidin Imidacloprid Thiamethoxam
lbs a.i./
100 lbs seed mg/seed
lbs a.i./
100 lbs seed mg/seed
lbs a.i./
100 lbs seed mg/seed
spinach 0.150 0.199 1.20
squash 0.750
sugar beet 0.720 0.893
sunflower 1.00 0.500 0.25
sweet corn 0.500 0.250 1.25
sweet potato 0.0100
taro 0.0100
turmeric 0.0100
wheat 0.070 0.094 0.052 Step 3: Calculating allowable application rates per planted acre
From the approximate neonic treatment rate per seed (Table 3), we can calculate the neonic application rate per acre of seeded
crop by looking at planting densities in use in California. These were obtained from a USEPA compilation (USEPA 2010).
California estimates were used where given in the compilation—otherwise, states in the Pacific Northwest were used as much
as possible. While seed treatment product labels sometimes carry an equivalent application rate per acre, we believe that rates
calculated here are more realistic. Given that an increasing proportion of growers do not actually know whether or not their
seeds were treated with neonics (Hitaj et al. 2020), it stands to reason that actual application rates per acre are driven by the
planting density, as chosen by the grower based on prevailing agronomics and their experience with the crop—rather than
seed treatment product labels.
In order to work out planting densities, the following measurements are needed: either the finished number of seeds to the
acre (when available); or a combination of the pounds of seed to the acre and a measure of seed weight—usually expressed
as the number of seeds per pound. Selected high and low estimates for seed weight and planting density compiled by USEPA
(2010) are reproduced in Table 4. They reflect different varieties of different crops.
Table 4. Seed weight and planting density extracted from USEPA (2010).a
Crop or crop group Seeds per acre
Seeds per acre
Lbs seed
per acre
Lbs seed
per acre
Seeds per
lb (low)
Seeds per lb
alfalfa 2,985,000 3,405,000 15.00 15.00 199,000 227,000
artichokes 2,722 2,722 3,400 3,400
Asian vegetables 142,362 142,362 0.44 0.44 325,400 325,400
asparagus 29,040 104,544 8.00 10.00
barley 282,000 420,000 30.00 98.00 9,400 14,000
beans, dry 69,696 104,544 38.42 130.68 800 1,814
beans, fresh 139,392 418,176 58.08 435.60 960 2,400
Table 4. Seed weight and planting density extracted from USEPA (2010).a
Crop or crop group Seeds per acre
Seeds per acre
Lbs seed
per acre
Lbs seed
per acre
Seeds per
lb (low)
Seeds per lb
beets 52,272 2,090,880 2.17 86.63 24,136 24,136
broccoli 62,726 69,696 0.42 0.87 80,000 150,000
Brussels sprouts 20,908 27,878 0.11 0.44 64,000 192,000
cabbage 22,402 26,136 0.14 0.58 45,000 165,000
carrots 900,000 1,300,000 2.25 7.43 175,000 400,000
cauliflower 17,424 23,232 0.12 0.29 80,000 150,000
celery 34,848 69,696 0.03 0.07 1,000,000 1,152,000
cereal, mixed 1,300,000 1,500,000 60.00 156.00 8,000 18,000
chard 13,068 17,424
chicory 29,040 52,272 0.08 0.14 377,320 377,320
cilantro 313,632 896,091
corn, grain 26,400 40,250 13.20 29.57 1,361 2,000
corn, silage 26,400 40,250 13.20 29.57 1,361 2,000
cotton 30,000 85,000 18.89 4,500
cucumbers 7,260 21,780 0.40 1.82 12,000 18,144
daikon 63,360 95,040
eggplant 6,534 14,520
escarole & endive 29,040 41,818
garlic 156,816 241,255
ginger root
ginseng 90.00 100.00
greens 167,000 144,000
hay, alfalfa 2,985,000 3,405,000 15.00 199,000 227,000
hay, cereal 780,000 1,088,640 60.00 90.00 13,000 18,144
herbs 435,600 726,000 1.47 4.84 150,000 296,500
kale 21,780 576,000 0.15 5.76 100,000 144,000
leek 43,560 87,120
lettuce 157,000 157,000 0.31 0.39 400,000 500,000
melons 29,040 34,848 1.40 2.18 16,000 20,800
mint 34,848 78,408
oats 780,000 1,088,640 60.00 90.00 13,000 18,144
okra 17,424 40,209
Table 4. Seed weight and planting density extracted from USEPA (2010).a
Crop or crop group Seeds per acre
Seeds per acre
Lbs seed
per acre
Lbs seed
per acre
Seeds per
lb (low)
Seeds per lb
onions 522,720 784,080 4.02 7.84 100,000 130,000
parsley 435,600 726,000 1.47 4.84 150,000 296,500
peas 196,020 522,720 39.20 163.35 3,200 5,000
peppers 10,890 26,136 0.15 0.52 50,000 72,000
potatoes 10,500 23,100 2,100.00 2,700.00 5 11
pumpkins 871 7,260 0.14 4.54 1,600 6,400
radishes 1,045,440 1,045,440 20.91 32.67 32,000 50,000
rhubarb 3,630 7,260
rice 1,742,400 2,439,360 77.00 118.00 15,600 28,100
rye 1,080,000 1,080,000 60.00 90.00 18,000
safflower 408,240 408,240 30.00 35.00 13,608
sorghum 45,000 100,000 0.66 9.09 11,000 68,040
spinach 360,000 408,240 9.00 25.00 40,000 45,360
squash 8,712 11,616 1.36 6.05 1,920 6,400
sugar beet 52,272 104,544 1.31 4.75 22,000 40,000
sunflower 6,000 27,000 3.00 4.00 2,000 9,000
sweet corn 15,682 59,739 3.48 33.19 1,800 4,500
sweet potatoes 11,880 18,341
tomatoes 52,272 69,696 0.28 0.58 120,000 190,000
turnips 29,040 261,360
vegetable seeds 52,272 69,696 0.28 0.58 120,000 190,000
wheat 1,300,000 1,500,000 60.00 156.00 8,000 18,000
wild rice
a Numbers in italics are calculated from the other entries in dierent columns of the table. Numbers in bold derive from sources other than the USEPA compilation.
Ranges obtained with calculated values tend to be wider than ranges given in the USEPA 2010 compendium. For example, the
calculated minimum of the range of pounds of seeds per acre is calculated by dividing the maximum number of seeds per acre
by the minimum number of seeds per pound, while the maximum pounds of seeds per acre is the minimum number of seeds
per acre divided by the maximum number of seeds per pound. This is probably a realistic estimate of the degree of variation
occurring in real life given different soil types and conditions as well as the myriad of cultivars for any given crop type;
whereas USEPA (2010) generally references a single source of information from the state—if any.
Combining the information from Tables 3 and 4 allows us to compute a neonic application rate, equivalent to seeding an acre
of any given crop. As mentioned earlier, maximum values are retained for the purpose of this assessment. These are presented
in Table 5.
Table 5. Equivalent application rates per acre for neonics in various crops with acreages compiled for this report. See text for
groupings and crop substitutions.
Crop or crop grouping
Clothianidin lbs
of a.i. per acre
Imidacloprid lbs
of a.i. per acre
Thiamethoxam lbs
of a.i. per acre
alfalfa 0.008
artichokes 8.01E-05
Asian vegetables 0.001 0.377
barley 0.069 0.092 0.051
beans, dry 0.173 0.163 0.065
beans, fresh 0.691 0.545 0.218
broccoli 0.326 0.015
Brussels sprouts 0.006
cabbage 0.122 0.006 0.006
carrots 0.344 0.143
cauliflower 0.005
celery 0.184
cereal, mixed 0.109 0.147 0.081
chard 0.046
chicory 0.006 0.138
corn, grain 0.111 0.156 0.111
corn, silage 0.111 0.156 0.111
cotton 0.079 0.095 0.070
cucumbers 0.036
escarole & endive 0.005 0.111
greens 1.080
hay, alfalfa 0.008
hay, cereal 0.063 0.085 0.046
herbs 0.080 1.921
kale 0.127
leek 0.069
lettuce 0.491 0.415
melons 0.058
oats 0.063 0.085 0.046
onions 0.553 0.346
parsley 0.080 1.921
peas 0.204 0.082
potatoes 0.338 0.338 0.162
Table 5. Equivalent application rates per acre for neonics in various crops with acreages compiled for this report. See text for
groupings and crop substitutions.
Crop or crop grouping
Clothianidin lbs
of a.i. per acre
Imidacloprid lbs
of a.i. per acre
Thiamethoxam lbs
of a.i. per acre
pumpkins 0.012
rhubarb 0.019
rice 0.166
rye 0.063 0.085 0.046
safflower 0.450 0.140
sorghum 0.018 0.023 0.021
spinach 0.135 0.050 1.080
squash 0.019
sugar beet 0.166 0.206 0.161
sunflower 0.040 0.015
sweet corn 0.066 0.083 0.165
sweet potatoes 0.338
vegetable seeds 0.019
wheat 0.109 0.147 0.081
wild rice 0.166 Step 4: Cross-referencing allowable application rates per acre with California agricultural
snapshot data
Steps 1 to 3 detailed above now allow for the calculation of potential total neonic use if all seeded acreages of registered
crops did in fact use a neonic seed treatment. Because rates vary among the three seed treatment active ingredients, several
assumptions were made regarding the relative market share of each active ingredient by crop. Where information was available
from USEPA sources (e.g., USEPA 2017, 2018, 2019a, c), this information was used to work out approximate market shares.
Where no information existed (e.g., as for all vegetable crops), all registered active ingredients were assumed to be used in
equal proportion. The initial approach had been to note the number of times specific crops were mentioned on labels and use
this as an indication of relative market share (see Appendix 2). However, it was decided that this was not a reliable indicator
and that the number of products labeled for any given crop had more to do with how different companies market their
products—some preferring to have separate labels for different crops rather than one “super-label” that tabulates all possible
crops and rates.
Crops for which no evidence of registered seed treatments could be found were excluded from the compilation. This included
asparagus, beets, daikon, eggplant, garlic, horseradish, okra, all pepper types, radishes, tomatoes, turnip, and watercress.
A few other crop-specific assumptions were made as follows:
To apportion market share of treatments in corn, we consulted the USGS published estimates of clothianidin,
imidacloprid, and thiamethoxam on corn for 2014, the last year when seed treatments were included in USGS use
estimates. For states bordering California (Arizona, Nevada, Oregon), the ratio of clothianidin to thiamethoxam use was
roughly 2:1 with no imidacloprid showing. This is the proportion we used for California.
For cotton, sugar beet, and wheat, the proportion of each neonic was assigned based on the country-wide information
provided in the analysis by USEPAs Biological and Economic Analysis Division or “BEAD” (USEPA 2017, 2018). The
proportional market share of the three active ingredients in wheat was extended to other cereal crops.
Clothianidin in sweet potatoes was assumed to be used at the same per acre rate as for potatoes, given the similarity
between the respective rates on seeds.
Table 6 provides the possible total quantity of each neonic active ingredient that could be applied via treated seed use,
assuming all seeds treated if allowed. County information was amalgamated by USDA agricultural region (see Figure 2).
Figure 2. Agricultural regions of California.
Table 6. The calculated potential use of neonics applied as seed treatments in 2016 by agricultural region of California.
Agricultural region
lbs of a.i.
lbs of a.i.
lbs of a.i. Total
Central coast 59,496 6,180 60,012 125,688
Northeast 899 5,849 4,693 11,441
Northern coast 40 357 221 619
Sacramento valley 7,386 30,261 24,551 62,198
San Joaquin valley 62,060 67,018 62,454 191,532
Sierra mountains 46 426 441 913
Siskiyou-Shasta 800 1,780 3,516 6,095
Southern California 34,321 12,815 47,590 94,726
Unspecifieda463 13,400 4,827 18,690
Total 165,511 138,086 208,305 511,902
a See section 2.2.1. e “Unspecied” total is based on USDA-reported seeded acreages that cannot be assigned to specic counties based on county statistics.
2.2.2. Comparing the total potential neonic-treated seed use to PUR data for non-seed neonic uses
We can compare the totals from Table 6 with those compiled by USGS from the PUR system in California (Table 7, Appendix
3). As mentioned earlier, the latter exclude seed treatments. Neither total includes consumer uses of these active ingredients.
Totals for the other neonics acetamiprid and dinotefuran are also included.
Table 7. Total application of neonics in California by certified applicators in 2016 as recorded in PUR system (lbs a.i.).a
region Clothianidin Imidacloprid Thiamethoxam
Sub-total of
the 3 principal
neonics Acetamiprid Dinotefuran
coast 7,998 40,906 12,807 61,711 6,056 1,969 69,736
Northeast 0 139 6 145 0 0 145
coast 0 516 8 524 2 0 526
valley 554 19,444 590 20,588 8,451 372 29,411
San Joaquin
valley 11,571 218,289 21,857 251,716 31,388 8,749 291,854
mountains 18 178 0 196 116 0 312
Shasta 4 208 26 237 7 0 245
California 1,977 66,787 6,445 75,210 4,242 2,323 81,775
Grand total 22,122 346,467 41,738 410,327 50,262 13,414 474,004
a PUR data compiled by USGS for 2016. Data obtained from:
As these calculations indicate, full deployment of nitroguanidine neonic seed treatments on crops for which the practice is
labeled exceeds the reported (PUR) use of the three active ingredients for 2016 (511,902 lbs a.i. vs. 410,327 lbs a.i.). Thus, not
taking seed treatment insecticides into account may seriously underestimate actual pesticide inputs in California.
Appendix 3 explores the potential discrepancy in more detail on a county by county basis. Reported use for counties reporting
0 lbs of neonic use was changed to 1 lb in order to calculate a proportional increase over the reported use. Clearly, those
counties with little to no PUR-reported neonic use, but with some crops allowing for seed treatment use showed the greatest
increase. However, even heavy agricultural counties showed substantial potential discrepancies suggesting that the use of
clothianidin and thiamethoxam, especially, may have been greatly underestimated. For example, the amount of thiamethoxam
used in Monterey county on the Central Coast could be over 500% of reported applications in the PUR system. Likewise, San
Joaquin County could have application levels 300% higher than reported.
2.3. Estimating the actual total neonic-treated seed use in California
There is no publicly available information on the actual use of neonic seed treatments in California. Even when reported
nationally, estimates are now very dated and not geographically explicit, and therefore potentially meaningless. Where
information on seed treatments has been tracked following registration (e.g., Douglas and Tooker 2015 for several field crops),
it shows that the extent of seed treatment use has increased exponentially. Later registrations and crops with smaller national
acreages have not benefitted from the same level of scrutiny. It is reasonable to expect that use patterns registered more
recently are still in their exponential growth phase. Also, it is anticipated that removal of the last uses of organophosphorus
insecticides (e.g., chlorpyrifos—see Appendix 4) will lead to further increases in neonic use.
Attempts to obtain seed treatment use information from California crop specialists yielded no relevant information. Crop
experts who were contacted and kindly responded did not know the use information, often recommending that I contact seed
retailers6 or suggesting (mistakenly) that the information could be obtained from PUR data. That total absence of knowledge
illustrates that, as predicted by Hitaj et al. (2020), failure to account for seed treatments does result in a loss of agricultural
Any information I was able to find is presented in Appendix 4.
This section provides a review of the potential for neonic seed treatment applications to contaminate state surface waters as
well as a short analysis of the neonic water sampling performed to date.
3.1. At what level do neonics cause ecological harm?
Before examining the water data in California, we must first examine at what levels neonics would be expected to cause harm
to aquatic ecosystems, also known as aquatic “benchmarks. An extensive review of the setting of aquatic benchmarks is
provided in my earlier reports on surface water contamination in New York State (Mineau 2019a, b). I will quickly summarize
the arguments here in the context of the California situation and explore recent developments. Relevant sections of these
reports will be excerpted here.
In my previous reports (Mineau and Palmer 2013, Mineau 2019a, b), I showed how USEPA had erred in its initial benchmark
setting exercises with imidacloprid and argued that they were now making the same error in setting their benchmarks for
clothianidin and thiamethoxam. My main thesis was that, given the 790,000 fold difference in sensitivity to imidacloprid from
the least to the most sensitive organism tested (with 36 species tested as of 2017), setting any benchmark based on the “most
sensitive” species for the smaller clothianidin and thiamethoxam datasets had more to do with chance than with good science.
As of June 2020, USEPA had derived the benchmarks appearing in Table 8.
Table 8. USEPA aquatic freshwater benchmarks in effect as of June 2020.a
Active ingredient Acute (µg/L) Chronic (µg/L)
Imidacloprid 0.385 0.01
Thiamethoxam 17.5 0.74
Clothianidin 11 0.05
a Data obtained from: (consulted June 2020).
Clearly, USEPA still believed then that clothianidin and thiamethoxam were safer to aquatic life despite ample evidence
that this difference was an artifact of the larger dataset for imidacloprid. In 2018, Canada’s PMRA (2018a, b) had already
proposed chronic benchmarks for clothianidin and thiamethoxam that were 33 and 28-fold lower than the proposed USEPA
benchmarks respectively.
As early as 2013, we (Mineau and Palmer 2013) had proposed that the aquatic toxicity of thiamethoxam and clothianidin
to aquatic insects and crustacea should be assumed to be similar to that of imidacloprid based on a comparison of toxicity
6 A current (Jan. 2019-June 2020) list of authorized dealers in California numbered more than 950.
tests performed on the same species with different neonics. The argument was strengthened and published in Morrissey et al.
(2015), which provided that:
In general, acute and chronic toxicity of the neonicotinoids varies greatly among aquatic arthropods. . . . Based on limited
data, however, it appears that differences in relative toxicity among the various individual neonicotinoids are minor.
(Morrissey et al. 2015)
Other authors have concluded as to the similar toxicity of nitroguanidine neonics (e.g., Hoyle and Code 2016) on the strength
of newer data such as Cavallaro et al. (2017). These authors generated comparative data for the three nitroguanidine neonics
on the same chironomid species. They found almost identical toxicities for imidacloprid and clothianidin—somewhat less for
The publication of more comparative data by Raby et al. (2018a, b) finally provided information sufficient to convince USEPA
that differences between neonic active ingredients were indeed not as great as it had originally believed (USEPA 2020c). The
upshot was that, for the more sensitive species tested, all nitroguanidine neonics should be considered to be of equivalent
When considering the toxicity data for the mayfly, all four chemicals are similar, with clothianidin, dinotefuran and
thiamethoxam all having 95% confidence intervals that overlap with the confidence intervals of imidacloprid. For the midge,
there are slight differences in toxicity among the chemicals, where both clothianidin and imidacloprid are similar (95%
confidence bounds overlap) and dinotefuran and thiamethoxam are slightly less toxic (LC50 values are 2x and 5x higher than
imidacloprid; confidence bounds do not overlap with those of imidacloprid or clothianidin).” (USEPA 2020c)
Similar results were obtained in the chronic toxicity tests with thiamethoxam being slightly less toxic than imidacloprid—but
by a two-fold difference only. It should be noted that thiamethoxam breaks down to clothianidin, so the lesser toxicity of the
former is not as relevant ecologically. No-effect concentrations for clothianidin and imidacloprid were within a factor of 4
and 2 for the most sensitive and second-most sensitive species respectively. Clothianidin was more toxic than imidacloprid
to the most sensitive species (a mayfly) but less toxic than imidacloprid for the second-most toxic, a chironomid. Clearly,
the differential toxicity ascribed to the different nitroguanidine products in previous USEPA aquatic risk assessments is not
justified scientifically.
Early research by Starner and Goh (2012) in California had shown that flowing water in agricultural watersheds showed
concentrations of imidacloprid that remained steady for periods exceeding three months at least. Similarly, Whiting et al.
(2014) and Whiting and Lydy (2015) had shown that, runoff water from corn seed treated with clothianidin carried residues
for the whole summer. Schaafsma et al.’s 2015 study indicated that residues could persist for a full year following a single
application. Given the persistence of the nitroguanidine neonics in soils or waters protected from direct sunlight, and the
fact that they are found to contaminate runoff water for months after application, it is clear that there is a very high potential
for chronic toxicity, making the chronic toxicity benchmark the most relevant threshold for ecological harm. Several authors
have commented on this (reviewed in Mineau 2019a, b). In addition, the cumulative toxicity potential of neonics has been
well examined. In a recent expansion of their previous analyses, Sanchez-Bayo and Tennekes (2020) restated their argument
that neonics show characteristics of irreversible cumulative toxicity in both aquatic and terrestrial invertebrates. Their sound
analysis argues for the fact that any benchmark based on acute or even short-term toxicity data is irrelevant in a real-world
exposure situation.
All of this argues for disregarding any acute benchmark in favor of a chronic one. In previous reports, I argued that the 0.01
µg/L chronic benchmark established by USEPA for imidacloprid should be applied to the other nitroguanidine seed treatment
chemicals. That has now been conclusively shown to be an appropriate benchmark for residues of any of the nitroguanidine
neonics in water, or, where more than one neonic chemical is present, the sum of all nitroguanidine neonic residues.
3.2. Predicted contamination from seed treatments relative to other
application methods
Because neonic seed treatments may be one of the greatest (if not the greatest) use of neonics in California, it is critical to
understand the extent of aquatic contamination one can expect from them. Two questions that arise are: (1) what is the
expected proportion of contamination from each chemical active ingredient; and (2) what is the expected contamination from
seed treatments relative to all other uses (soil, foliar, etc.).
On the first question, as discussed below, surface water testing for clothianidin and thiamethoxam—both of which are
commonly used as seed treatments—is inadequate. However, we can glean some insight from the active ingredients’ surface
water mobility indices (SWMIs)—a measure designed by Chen et al. (2002) for the mobility of a chemical in the environment,
and therefore its propensity to contaminate surface waters. The index ranges from 0 (low mobility) to 1 (extreme mobility),
and the SWMI value for each of the three main nitroguanidine neonic active ingredients is provided in Table 9.
Table 9. Surface Water Mobility Indices (SWMIs) for the main seed treatment neonics based on an algorithm designed
by Chen et al. (2002).a
Neonic active ingredient SWMI value
Clothianidin 0.66
Imidacloprid 0.56
Thiamethoxam 0.82
a Input data from Pesticide Properties Database obtained from:
As Table 9 shows, the physico-chemical properties of clothianidin and thiamethoxam make them more likely to contaminate
surface waters than imidacloprid. Therefore, given similar use patterns, the predicted surface water contamination from
clothianidin and thiamethoxam would be more extensive than for imidacloprid.
On the second question, we can first look to attempts by USEPA to assess the risks of neonic water contamination from seed
treatments versus other sources in their recent re-evaluation of neonic impacts, which took place as part of a process known
as “registration review. In their initial review of imidacloprid residues in surface waters in the United States at large, USEPA
(2016) used modeling to predict water concentrations from labeled uses. They argued that runoff concentrations from soil
and seed treatment uses would be lower than from foliar applications, as shown in Figure 3. (Seed EECs).7
Figure 3. Detected imidacloprid concentrations in U.S. surface waters from routine monitoring efforts relative to the range of
modeled concentrations for different use patterns of the pesticide (USEPA 2016).
7 EEC stands for “Expected Environmental Concentration,” an estimate generally obtained through models of runo and/or dri.
Figure 2 shows a great deal of variability in the expected water concentration following the use of seed treatments. Part of this
variation has to do with the labeled rate on seed as well as the depth of incorporation into the soil profile. The runoff models
used by USEPA initially predicted no runoff when seeds are planted at more than 2 cm in depth (USEPA 2016).8 Given that
recommended seeding depths for both corn and soybean are deeper than this 2 cm depth, the model findings were clearly at
odds with the fact that aquatic contamination is very extensive in areas planted to corn and soybeans (Hladik et al. 2014). This
contradiction was raised, but not resolved, in USEPA (2016).
USEPA (2020c) summarized the expected environmental concentrations (EECs) it anticipates from various seed treatment
uses with the three main seed treatment neonics (Table 10). These are the water concentrations USEPA used to calculate risk
quotients and assess the likely environmental damage from seed treatment uses in its recent re-evaluation. These estimates are
meant to be conservative (i.e., worst case estimates) at this stage of the evaluation process. However, as seen below, they are
clearly not.
Table 10. Estimated environmental water concentrations (EECs) in µg/L predicted by USEPA for neonic seed treatment
uses (USEPA 2020c).
Peak (1 day) EEC µg/L Chronic (21 day) EEC µg/L
Clothianidin Imidacloprid Thiamethoxam Clothianidin Imidacloprid Thiamethoxam
Cotton 2.14 2.19 1.18 1.87 1.56 0.97
Corn 0.59 0.78 0.39 0.53 0.56 0.33
Soybean 1.77 3.33 0.46 1.53 2.21 0.40
Sugar beet 1.11 2.63 1.35 0.96 1.83 1.24
Wheat 2.27 2.56 1.08 1.99 1.81 0.86
Rice 71.7 NA 66.4 8.74 NA 35.5
Indeed, long before USEPA published these estimates, there was already evidence from published field studies that the agency
had underestimated the potential for seed treatments to contaminate surface water. One possible reason for this is that their
modeling ignores the issue of dust that is produced at seeding (USEPA 2016).
In the independent literature, Main et al. (2014) reported clothianidin values as high as 3.1 µg/L from sloughs (small ponds
that occur frequently in “knob and kettle” landscapes) in canola-growing areas following the use of clothianidin seed
treatments; and Schaafsma et al. (2015) reported levels as high as 16.2 µg/L in ditches outside another corn field seeded to
clothianidin and 3.25 µg/L in puddles as far as 100 m from the fields. Whiting et al. (2014) and Whiting and Lydy (2015)
documented clothianidin residues of 0.23 µg/L in runoff from a corn field, but this was at one-fifth of the allowable treatment
rate;9 more importantly, residues persisted in runoff water a full 156 days after planting.
The situation is similar with thiamethoxam seed treatments. Main et al. (2014) found values up to 1.49 µg/L from sloughs
around canola fields; and Schaafsma et al. (2015) measured levels as high as 7.5 µg/L in ditches outside a seeded field and 16.5
µg/L in puddles outside their Ontario corn fields. The latter two measurements were even more remarkable because they were
measured pre-plant and therefore indicated contamination from the previous use of seed treatments in the preceding growing
season. Higher levels were recorded in puddles within the field area.
More in line with USEPA (2020c) predictions, recent samples taken from a variety of waterbodies in crop and non-crop sites
within an agricultural landscape in Indiana (Miles et al. 2017; with 2018 correction) found concentrations of clothianidin
averaging 0.101 µg/L (all sites combined; with samples taken weekly for eight weeks). The highest concentrations of
8 ese predictions, obtained through the PWC (Pesticide in Water Calculator), provide 2 cm as the default value beyond which no runo is expected, although this input
can be changed by the user. Data obtained from:
(consulted June 2020).
9 It is common practice to scale contamination levels with application rate. In this case, this would result in a contamination level approximating 1.15 µg/L.
clothianidin (0.45-0.67 µg/L) were observed in small lentic woodland bodies of water well away from the seeded corn
and soybean fields. One of these sites (PWA West) apparently received drainage from nearby fields; how the other got
contaminated is unknown.10 Regardless, levels in these wetlands were higher than those reported in any of the ditch samples
taken nearer the seeded fields, showing the difficulty in containing these extremely mobile insecticides within the treatment
On the whole, there is now ample model and field data evidence predicting what the expected neonic water contamination
would be from seed treatment use. Regardless of whether current USEPA model estimates represent an accurate picture or a
considerable underestimate—as the independent scientific literature would suggest—the anticipated neonic levels in water
are all well within the range where we would expect considerable injurious impacts on aquatic habitats receiving runoff.
Therefore, even if expected contamination levels from seed treatments are lower than from other application methods, the
potential scale of use (as outlined in Section 2) is a clear concern with respect to the California environment.
3.3. Evidence of neonic contamination in California surface
water data and likely environmental impacts
In California, CaDPR amalgamates all known water analyses into the state Surface Water Database or “SURF.11 The database
includes CaDPR studies as well as the national USGS water sampling programs and others. CaDPR has been analyzing surface
waters for pesticide residues since 1981.12 Since 2000, the CaDPR program performing these analyses has been the Surface
Water Protection Program (SWPP) (Goh et al. 2019).
Knowledge of extensive contamination of California surface waters by imidacloprid is not new. In 2012, Starner and Goh
presented data from 2010-2011 demonstrating such contamination. Some of the sampling sites used in that report have
continued to be monitored over time, and data from those sites are presented below.
Hoyle and Code (2016) queried SURF for the presence of neonics in samples collected between January 2010 and October
2015. For imidacloprid, this amounted to 790 samples taken from 132 sample sites. They found that 55% of the sites had at
least one imidacloprid concentration above the level of detection with a mean detection level of 0.643 µg/L. Despite using the
older EPA benchmark of 1.05 µg/L (105X higher than the currently accepted benchmark), they found that 14% of samples
fell within the range expected to cause significant biological effects on receiving waters. Hoyle and Code (2016) further
pointed out that, using the more protective European chronic benchmark of 0.067 µg/L (closer to the current USEPA and our
proposed 0.01 µg/L benchmark), 89% of imidacloprid detections exceeded this benchmark.
Neonic use has dramatically increased since the 2010-2015 period when these water samples were taken. Hoyle and Code
(2016) only examined imidacloprid data and, as they point out, it is difficult to separate the agricultural uses of imidacloprid
from its domestic and landscaping uses, although high rates of detection in the Santa Maria, Salinas, and Imperial Valley areas
did suggest agriculture is an important contributor.
Rather than repeating the Hoyle and Code (2016) analysis, I approached the data differently. Without an analysis of each of
the sampling locations, their flow rates, why sites were chosen, and how samples were timed to correspond to agricultural
activity and rainfall, the proportion of positive imidacloprid detections does not paint an accurate picture of whether they
reflect agricultural or urban sources. New data collected by CaDPR, however, does address this to a certain extent (see below).
Data on clothianidin and thiamethoxam—both of which are almost entirely associated with agriculture—are also now
The SURF database was queried in May 2020. It consisted of results tabulated to December 2019. Samples consisted of either
filtered water samples or whole water samples, but this was not specified in all cases. Because of the high solubility of the three
neonics of interest, I ignored this parameter, estimating minimal loss of analytes from filtering out the particulates.
As mentioned, CaDPR monitors both agricultural and urban sources of imidacloprid contamination. Studies are identified
as such in the SURF database. I assumed that, if designated as such, the chosen sample locations indeed reflect ongoing
10 is is based on inspection of the PWA East & West sampling site photographs included in the publication and personal communication with two of the authors:
C. Krupke and J.T. Hoverman.
11 Data obtained from: (consulted June 2020).
12 Data obtained from: (consulted June 2020).
agricultural or urban land use, although we cannot say whether samples were chosen to represent maximum likelihood of
finding the analyte—e.g., presence of cropped fields or time of sampling relative to use patterns and precipitation, type of
waterbody, etc. As reviewed in earlier analyses (Mineau 2019a, b) samples taken as part of regular water monitoring programs
always underestimate the presence and levels of targeted pesticides.
Over the entire 2010-2018 period,13 CaDPR sampling for imidacloprid in sites labeled “agricultural” was comprised of
556 samples from 86 sampling sites (Table 11). Imidacloprid was detected at half of those sites, indicating some use in the
watershed. All sampled sites had maximum detection levels that exceeded the 0.01 µg/L benchmark and a large proportion
exceeded the benchmark by 10X or even 100X (Table 11). Indeed, over half of the sampled sites had maxima that exceeded the
USEPA acute benchmark of 0.38 µg/L, a level which, as I argued earlier (see section 3.1), is clearly not protective enough in the
case of a persistent compound with cumulative toxicity. The highest recorded maximum level was 41.1 µg/L. The next highest
was obtained on the same site, suggesting the maximum reading was not in error. It came in at 9.86 µg/L.
The highest imidacloprid levels were recorded in Monterey County, followed by Santa Barbara, Imperial, and Napa—all
intensive agricultural regions. In these counties, the imidacloprid data alone clearly indicates serious adverse impacts for
aquatic life in receiving waters. Looking at positive detections over the years (Figure 4), there appears to be a trend for higher
levels of detection over time. This could be the result of more targeted monitoring or reflective of the generally increasing
amount of use of imidacloprid over that period.
CaDPR also sampled urban sites throughout the same period. A total of 578 samples were noted as being part of urban
monitoring studies—roughly half from Orange County (Table 11). Although more of the sampled sites had positive
detections, impacts from urban use may be less pronounced than for agricultural uses, as judged by the fewer number of sites
registering maximum levels either 10X or 100X benchmark levels. Looking at positive detections over the years (Figure 5)—
with the exclusion of one extreme value for 2018—does not reveal any convincing pattern, although there is a slight suggestion
of a decline. Given that several of CaDPR’s studies mention “mitigation monitoring,” one can assume that there has been some
effort at containing urban sources of imidacloprid. However, the generally lower levels seen in the urban samples may also
be a result of higher flow rates and larger bodies of water being sampled. The ubiquitous presence of neonics in storm water
retention ponds and canals has been of increasing concern on the part of municipal and regional governments (Murray 2015).
As was the case with the agricultural samples, one site had an extreme value (165 µg/L) recorded in 2018. The next highest
detection (12.7 µg/L) was recorded at the same site some years earlier.
Table 11. CaDPR monitoring results for imidacloprid in studies indicated as agricultural or urban in nature.
Study type
All sites Sites with positive detections
of samples
of sites
Number of
sites with
level (ppb)
Sites with
Sites with
10X above
Sites with
100X above
Agricultural 556 86 43 (50%)
43 (100%) 33 (77%) 15 (35%)
Urban 578 57 43 (75%)
43 (100%) 21 (49%) 3 (7%)
13 As of May 2020, no CaDPR agricultural monitoring samples were reported beyond 2018.
Figure 4. Plot of positive detections (in µg/L) by year for CaDPR’s agricultural sampling for imidacloprid. The outlying value
of 40.1 µg/L from 2017 was omitted. The overall trend in the mean is depicted by the red line.
Figure 5. Plot of positive detections (in µg/L) by year for CaDPR’s urban sampling for imidacloprid. The outlying value of 165
µg/L from 2018 was omitted. The overall trend in the mean is depicted by the red line.
3.3.1. Continued monitoring of imidacloprid at sites from the Starner and Goh (2012) study
As mentioned above, the California data from agricultural sites reported by Starner and Goh (2012) offered good evidence
that receiving surface waters were exposed to residues over the entire growing season, which argues in favor of using a chronic
benchmark as the ecologically relevant one.
Monitoring continued at many of those sites right up to 2018. I chose to show data from two of the most intensively sampled
sites—two tributaries of the Salinas River in Monterey County, Chualar Creek, (Figure 6) and Quail Creek (Figure 7).
Figure 6. Imidacloprid residues at site 27-8 – Chualar Creek, a tributary to the Salinas River, Monterey County. The scale is
shown on a log scale in order to accommodate the wide range in values.
In the case of Chualar Creek, every single water sample taken at that sampling site since 2010 has had a concentration of
imidacloprid over 10X the benchmark level for damage to aquatic life (i.e., over 0.1 µg/L) and frequently over 100X that
benchmark (i.e., over 1.0 µg/L). The final sample taken in 2018 was the highest at 41.1 µg/L.
Quail Creek shows a similar, although not quite as extreme, pattern. Over the 8 years of sampling, water concentrations
measured between May and November seldom dipped below 0.5 µg/L or 50 times over the benchmark level for damage to
aquatic life.
Figure 7. Imidacloprid residues at site 27-7 (Quail creek), a tributary to the Salinas River, Monterey County.
3.3.2. Clothianidin and thiamethoxam sampling
Sampling for clothianidin and thiamethoxam in California waters has not been adequate. Although some sampling for
clothianidin and thiamethoxam began as early as 2011, it is clear that much of the sampling locations, especially in the earlier
years, were not chosen with agriculture in mind, but as part of on-going USGS sampling of urban/industrial areas—e.g.,
Mallard Slough in Sacramento, or estuarine areas such as Grizzly Bay. Later, other groups became involved in sampling, such
as the California State Water Resources Control Board’s Surface Water Ambient Monitoring Program or “SWAMP, as well
as other coalitions of different regional water boards. Fortunately, these sampling efforts include some of the same CaDPR
agricultural long-term monitoring sites, as CaDPR by 2018 had still not incorporated clothianidin or thiamethoxam into its
agricultural sampling programs. However, these samples were not taken during the more intensive summer season.
As a result, of the 85 sites monitored for imidacloprid by CaDPR, only 11 (13%) were examined for clothianidin or
thiamethoxam residues, and this was only for the years 2017-2018.
Table 12 compares levels of imidacloprid with those of clothianidin and thiamethoxam for the 11 CaDPR agricultural
monitoring sites. As mentioned, the comparison is rather inadequate because all of the clothianidin and thiamethoxam
samples were taken in either September or December, whereas imidacloprid samples for the most part, were taken between
April and September. As discussed in Mineau (2019a, b) the higher the frequency of sampling and the higher the number
of samples, the greater the chance of detecting the higher residue concentrations. For example, the “maximum” value of
0.398 µg/L thiamethoxam obtained after only two winter samples at site 27-66 compared to the maximum 0.647 µg/L of
imidacloprid following 47 spring and summer samples of imidacloprid strongly suggests that thiamethoxam would be shown
to be a more significant contaminant at that site were the sampling adequate.
Table 12. Levels of three neonicotinoids compared at agricultural monitoring sites sampled by CaDPR. Samples taken from 2010-
2018 for imidacloprid, and 2017-2018 for clothianidin and thiamethoxam.
Site Clothianidin Imidacloprid Thiamethoxam
Maximum recorded
concentration (µg/L) N
Maximum recorded
concentration (µg/L) N
Maximum recorded
concentration (µg/L)
Monterey 27-12 2 0.060 36 1.12 2 0.482
Monterey 27-13 1 0 27 0.067 1 0.080
Monterey 27-2 1 0 1 0 1 0
Monterey 27-50 2 0 8 0.167 2 0.353
Monterey 27-58 1 0.251 5 0.118 1 1.44
Monterey 27-66 2 0 47 0.647 2 0.398
San Luis
Obispo 40-13 2 0 26 1.44 2 0.069
San Luis
Obispo 40-23 2 0 6 0 2 0.071
Santa Barbara 42-48 2 0 26 9.14 2 0.113
Santa Barbara 42-49 2 0 11 1.39 2 0.001
Santa Barbara 42-50 2 0.046 34 4.91 2 0.427
Santa Cruz 44-18 2 0 4 0.064 2 0.040
a N denotes the number of samples reported.
These data indicate that clothianidin and, especially, thiamethoxam are major contaminants of agricultural areas of
California—which is perhaps predictable given their physico-chemical characteristics (Section 3.2) as well as the large
reported and potentially larger unreported use (Section 2.3). Despite very fragmentary and inadequate sampling, maximum
thiamethoxam water levels have already surpassed those of imidacloprid at some sites. Given that the potential clothianidin
and thiamethoxam seed treatment use may far exceed their other uses in many counties of California (see Appendix 3), the
failure to include these insecticides within a comprehensive agricultural sampling program or track seed treatment products
more generally, has created a serious data gap—depriving scientists and policy makers of key information about the likely
significant and damaging impacts of neonic-treated seed use.
Table 13 shows the total number of positive clothianidin and thiamethoxam detections regardless of sampling site. As
discussed earlier, sampling intensity was much lower than for imidacloprid, so benchmark exceedances are not strictly
comparable—especially since many sites are not in agricultural areas and clothianidin and thiamethoxam are predominantly
used in agriculture.
Table 13. Positive detection of clothianidin and thiamethoxam in California (2011-2018).
Clothianidin Thiamethoxam
Number of sample sites with positive samples 23 47
Maximum level recorded (µg/L) 1.6 9.0
Sites with detections above 0.1 µg/L benchmark 20 (87%) 38 (81%)
Sites with detections 10X above 0.1 µg/L benchmark 13 (57%) 25 (53%)
Sites with detections 100X above 0.1 µg/L benchmark 4 (17%) 9 (19%)
The fact that each active ingredient, despite the inadequate sampling, frequently exceeds benchmark levels is worrisome. Given
that most aquatic systems with agricultural regions will be exposed to these three nitroguanidines—along with the other
neonics dinotefuran and acetamiprid—in combination, their combined concentrations must be considered when assessing
injury to the aquatic environment. Clearly, not sampling comprehensively for neonics other than imidacloprid misses a
critically important part of the picture.
4.1. Recent regulatory reviews of neonics and pollinators
USEPA and CaDPR collaborated on an extensive review and risk assessment of the three main nitroguanidine neonic
insecticides (clothianidin, imidacloprid, and thiamethoxam) regarding their impacts on bees.14 The regulatory agencies, in
conjunction with the registrants, developed protocols for honey bee colony studies to be carried out by the registrants. These
studies formed the core of the assessments. Given the sheer amount of information available to these agencies, as well as the
multitude of neonic use patterns, this evaluation was a complex one. CaDPR (CaDPR 2018, 2019) was first in publishing
its version of the assessment, the review document and addendum coming in at almost 1200 pages. The USEPA final
published version of the assessment (USEPA 2020d, e, f, g, h, i, j) came in at over 1100 pages, once the various appendices and
attachments are factored in. A detailed review of either the USEPA or CaDPR versions of the pollinator assessment is beyond
the scope of this report; but this report will highlight a few notable differences between the two assessments as well as some
clear inadequacies and shortfalls, specifically with respect to the assessment of seed treatment uses. Simply put, the CaDPR
and USEPA assessments both failed to adequately characterize the risk of neonic seed treatments to bees, especially wild bees.
This is a serious issue because, while other methods of neonic application may indeed carry a higher risk on a per-application
basis, the use of seed treatments can greatly extend the risk to pollinators spatially and involve crops not yet assessed for bee
With the methods outlined in Chapter 2 and Appendix 1, this report can account for a little over 5 million acres of non-
organic cropland in the State of California. As reviewed earlier, a number of minor crops—including several for which seed
treatments are allowed—are not included in this total. Nevertheless, we can calculate that 76% of this crop area can potentially
be sown with a neonic-treated seed. On a county by county basis (see Appendix 5), the proportion of crop area potentially
seeded with neonics ranges from a low of 17% (Yuba County) to 100% (Imperial County). The sheer size of the potential
treated area is such that it is critical to adequately assess the risk to pollinators posed by seed treatments.
4.2. What does the CaDPR assessment conclude with respect to neonic-
treated seeds?
CaDPR (2018) essentially dismissed any risk from neonic-treated seeds on the basis of preliminary joint assessments carried
out by itself and with USEPA. On the basis of these preliminary efforts, CaDPR concluded that residues in pollen and nectar
resulting from seed treatment use were either below detection limits or at least below any effect levels. As detailed below, the
USEPA assessment is much more nuanced and does, in fact, indicate a measurable risk to both individual pollinators and
honey bee colonies, even before errors of omission and commission in that assessment are taken into account.
CaDPR (2018) does raise the issue of abraded seed coat dust (see extensive review below) but concludes that USEPA has
addressed the issue with best management practices and that there are no records of such incidents occurring in California. To
the author’s knowledge, the issue of dust at seeding might indeed have been partially addressed, but it is a long way from being
CaDPR only considers risk to honey bee colonies and therefore does not look at risk to individual pollinators in the field.
The risk determinations are based on deriving no observable adverse effect concentrations (NOAEC) for pollen and
14 In fact, Canada’s PMRA also contributed in this tripartite eort.
nectar residues in treated crops from registrant-supplied honey bee colony feeding studies (Table 14). In other words, the
concentrations under which no effects were observable on honey bee colony health.
Table 14. NOAEC values in nectar and pollen used by CaDPR to assess the risk of residues in various crops (modified
from Table 1; CaDPR 2018).
Active ingredient NOAEC in nectar (ng/g) NOAEC in pollen (ng/g)
Clothianidin 19 372
Imidacloprid 23 97.5
Thiamethoxam 30 372
One notable difference between the CaDPR and USEPA assessments is that CaDPR generates separate risk ratios for nectar
and pollen consumption. The CaDPR assessment compares nectar and pollen residue levels in crop plants to the NOAEC
levels generated in separate whole hive studies conducted with either spiked sugar solutions (surrogate for nectar) or spiked
pollen patties. USEPA uses a “total dietary approach” and makes the assumption that honey bees will consume 20X more
nectar than pollen; pollen residue concentrations are therefore divided by 20 to derive “nectar equivalent” residue values. As
discussed below, neither approach is satisfactory for assessing the risk to wild bees.
By not considering risk at the individual bee level, CaDPR completely excludes appropriate consideration of any of the 1500+
Californian wild bee species, especially solitary bees, where the death of individuals is the appropriate endpoint against which
potential exposure should be compared. In addition, there are many independent studies in the literature that show effects
on bumble bee and mason bee colonies at lower exposure concentrations (see below). Curiously CaDPR offers that: Apis
bees serve as a surrogate for other non-Apis species of bees (e.g., bumble bees) that may be exposed under agricultural conditions
(CaDPR 2018). In actual fact, honeybees are not adequate surrogates and a honey bee-centric exposure scenario does not
inform on the risk to wild bees as argued below. By considering only honey bee colony studies in its assessment, CaDPR
explicitly excludes any consideration of solitary bee species.
In the rest of this chapter, we will look at the fuller evaluation of seed treatment uses in the USEPA assessment.
4.3. What does the USEPA assessment conclude with respect to neonic-
treated seeds?
The USEPA assessment finds that, for imidacloprid, risk to individual bees falls below levels of concern for all crops except
beans, canola, cotton, peanuts, peas, safflower, soybeans, and sunflower. The highest risks are for beans and peanuts, where
honey bee colony-level risks of concern were identified. However, the strength of evidence supporting this risk finding was
considered “weak.
For clothianidin and thiamethoxam, the assessment concludes that the risk from seed treatments to honey bee colonies was
low for all uses except clothianidin application to turmeric seed pieces, but the strength of evidence was considered to be weak
there also. However, risk to individual bees was above the level of concern for canola, corn, legumes, sorghum, and soybean
(for clothianidin) and beans, cucurbits, legumes, lentils, peanuts, peas, sorghum, soybeans, and sunflower (for thiamethoxam).
It is useful to recall that the earliest reported honey bee toxicity incidents with imidacloprid resulted from its use as a seed
treatment in corn and sunflowers (Bonmatin 2005, Maxim and van der Sluijs 2013). As early as 1994, reports were coming in
describing aberrant bee behavior, loss of foragers, and reductions in honey production, especially near sunflower crops. The
first regulatory restrictions in France, in 2004, were for these seed treatment uses in response to a strong lobbying effort from
the bee industry. Indeed, many honey bee mortality incidents reported to USEPA resulted from seed treatment uses.
It is therefore reasonable to ask whether the USEPA assessment is correct in dismissing most risks associated with seed
treatment uses. The rest of this chapter explores this question, although some of the issues and problems with the assessment
of seed treatments clearly have a bearing on the assessment of other neonic use patterns.
4.4. Deficiencies in USEPA’s assessment of pollinator risk from neonic-
treated seeds
4.4.1. The USEPA assessment fails to adequately consider risks to pollinators beyond honey bees
Honey bees (Apis mellifera) are important for the pollination of certain crops in California and, because of that economic
imperative, the USEPA assessment like the California DPR assessment is very “Apis-centric. Other species are mentioned in
passing and a few bumble and mason bee studies are reviewed, but the emphasis of the assessment is on honey bees.
Like CaDPR, USEPA (2020d-j) concludes that honey bees can be used as an adequate surrogate for non-Apis bees—and
this because of the circular logic that “… the bee risk assessment framework used by the EPA indicates the honey bees are
intended to be reasonable surrogates for other bee species …” (USEPA 2020e; section 6.2) (emphasis added). This “intention
notwithstanding, the USEPA assessment offers plenty of evidence to the contrary; some of this evidence will be highlighted
below. Wild bees may be more sensitive than honey bees
It is known that toxicity within and among bee species can vary greatly. For imidacloprid, USEPA (2020d) reports that within
honey bees alone, the acute oral LD50 values in tests they deemed acceptable ranged a full two orders of magnitude (100X).
Taking the lower value (as was done in the assessment) may provide a certain amount of added safety, but it isn’t clear whether
this is sufficient to protect the many species of wild bees that will be exposed in California cropland. Indeed, several authors
have looked at the relative sensitivity of bee species to a wide range of insecticides (providing more data than what is available
for neonics only) and concluded that honey bees are not always the most sensitive for any given active ingredient.
The USEPA assessment also references studies that show that leafcutter bee larvae are more sensitive than honey bee larvae to
clothianidin, that adults of the only stingless bee tested (Nannotrigona perilampoides) may be more sensitive than honey bees
to thiamethoxam, and that bumble bees may be more sensitive to clothianidin and/or thiamethoxam through the oral route.
In their proposed methodology for pollinator assessment, the European pesticide regulators (EFSA 2013) proposed an
extrapolation factor of 10X from the accepted honey bee LD50 in order to cover other species. This was, in part, a result of a
literature review comparing a handful of species only, so they expressed the view that this factor was preliminary and may be
insufficient. A more recent analysis of paired toxicity data from the same studies (Arena and Sgolastra 2014) obtained a similar
result: a safety factor of 10X applied to the honey bee toxicity endpoints was sufficiently protective of other bee species in 95%
of cases. However, the full range of sensitivity ratios between the honey bee and one of the other 19 bee species with which it
was paired ranged over 6 orders of magnitude. The differential weight of test bee species appears to be part of the reason for
this vast difference, but other factors are also at play (Arena and Sgolastra 2014).
Accordingly, the assumption that the honey bee will be more sensitive that 1500+ wild bee species potentially exposed in
the California cropland and adjacent non-crop areas is highly unrealistic. Moreover, a honey bee colony contains tens of
thousands of bees; many more individuals can be lost or affected before the hive is compromised. This alone makes the
conclusions of the USEPA assessment highly suspect given that the final risk estimate is based on honey bee colony-wide
endpoints. For solitary bee species or bees with smaller colonies (e.g., bumble bees) the loss and/or debilitation of individual
foraging adults becomes much more critical to population survival. Also, as seen below, the parameters of the USEPA
assessment—e.g., the low pollen ingestion by individual honey bees—do not adequately represent other species.
As an aside, several studies reviewed in the USEPA assessment have provided contact LD50 estimates that are lower than the
ones finally adopted by USEPA. These studies were found to be acceptable but labeled “qualitative” because of the lack of
access to the raw data. Given the very small difference obtained with different probit methodologies (i.e., how raw data are
processed to arrive at an LD50), this data exclusion seems excessive. Had the non-industry data been accepted, it would have
reduced the contact LD50 by over 3X. The estimates above also ignore that breakdown products of imidacloprid, especially
(imidacloprid olefin), which appear to be more toxic than the parent material (e.g., Suchail et al. 2001).
Yet another serious limitation of the USEPA risk assessment is that it has failed to also consider that the presence of fungicides
often included on the same seed treatment, as neonics can greatly enhance the toxicity of the latter. For example, Tsvetkov et
al. (2017) showed that clothianidin and thiamethoxam both became twice as toxic in the presence of boscalid at relevant field
15 is was established by measuring levels in collected pollen.
NEONIC SEED TREATMENTS IN CALIFORNIA | 29 Wild bees may be in different parts of the agricultural environment than honey bees
The USEPA assessment does concede the fact that many species of wild bees are attracted to crops that honey bees are not
attracted to. In addition, many different matrices (both biotic and abiotic, such as foliage or mud) are potentially attractive to
non-Apis bees. The USEPA assessment only looks at pollen and nectar as sources of exposure. Hopwood et al. (2016) estimates
that 70% of North American bees are ground nesting. According to the same authors, most solitary bee species have a smaller
foraging range and higher crop fidelity than honey bees. In the USEPA assessment, true cereal crops (for example) are deemed
not to be attractive to honey bees and therefore to carry no risk. However, this ignores the possibility of arable weeds in
flower in the crop, as well as movement of residues away from treated areas, either through dust drift or runoff. This off-site
movement is reviewed in section 4.5 below. By excluding any consideration of crops deemed not to be attractive to honey
bees, or crops harvested before flowering, USEPA seriously underestimates potential impacts to both honey bees, and wild bee
species. The USEPA risk assessment for individual honey bees is inadequate for wild bee species
To assess the risk to honey bees, the USEPA assessment uses the endpoints summarized in Table 15.
Table 15. Toxicity endpoints used in the USEPA assessment—said to be the lowest endpoints for which raw data were
available to allow independent statistical verification—modified from Table 1-2 in USEPA (2020d) and 1.3 in USEPA
Study type Assessment endpoint Clothianidin Imidacloprid Thiamethoxam
Adult acute contact 96-hour LD50
µg a.i./bee 0.0275 0.043 0.021
Adult acute oral 48-hour LD50
µg a.i./bee 0.0037 0.0039 0.0038
Adult chronic
(10-day) oral
µg a.i./bee/day
(12% mortality)
(food consumption)
(70% mortality)
Larval chronic repeat
dosing (21-day)
µg a.i./larva N/A 0.0018/>0.0018 0.0037/0.0066
Colony feeding study
(Spiked sucrose
(ng/g) 19/36 23/47 44/82b
a Values for clothianidin and thiamethoxam are expressed as clothianidin-equivalents (thiamethoxam values being corrected based on the ratio of their molecular weights).
b When assessing risk to honey bee colonies from thiamethoxam, USEPA uses the lower clothianidin values of 19 NOAEC and 36 LOAEC in acknowledgment of the fact that
thiamethoxam is transformed to clothianidin in crop plants and in the environment.
For reasons mentioned above, it is appropriate to consider which endpoint should be used in the first tier of the risk
assessment, which considers risk to individual foraging bees.16 In its “refined Tier 1 assessment” (i.e., assessment of risk to
individual bees based on measured pollen and nectar residue levels) USEPA assessment computes risk estimates for both acute
and chronic exposure scenarios. For a seed treatment use and given the duration of flowering in most crops, it could easily be
argued that the chronic (10-day) toxicity test is much more relevant ecologically, especially in the case of bee species with a
more limited range.
Because of the contamination of both nectar and pollen, the oral toxicity endpoints are arguably the most relevant for looking
at potential effects from seed treatments, once the initial “dust deposition” (see below) has passed. The oral toxicity of all three
active ingredients is shown to be virtually identical in the 48-hour adult oral test. Given the variability in test results seen
for any given active ingredient (discussed above), this similarity is quite remarkable. However, the lack of a consistent ratio
between the 48-hour values and the 10-day values among the different pesticides may have to do more with the vagaries of
16 It has been argued above that this endpoint should be 10X below that established for honey bees in order to cover all bee species, but we will leave this line of argumentation
aside for now.
testing (test conditions, bee race and provenance, bee age, etc.) than with any real toxicological or pharmacological difference
between the active ingredients. As reviewed by Sanchez-Bayo and Tennekes (2020), both clothianidin and thiamethoxam show
a similar relationship between toxicity and exposure time in honey bees. Because of the negligible amount of metabolism,
toxicity is effectively cumulative, so the clothianidin chronic results are exactly in line with what one would expect—the lethal
dose per day over a 10-day period is approximately one-tenth the one-time lethal dose. Thus, EPAs thiamethoxam chronic
oral value (see Table 15 above), being almost the same as the one-time oral dose value, is the aberrant one and not credible.
Accepting for the time being that 0.00036 µg/bee is a more ecologically realistic endpoint for the toxicity of the nitroguanidine
neonics, we can back-calculate the nectar concentration needed to deliver this dose to a foraging bee over a 10-day period.
Deriving “acceptable” or “limit” concentration levels in nectar and pollen is a direct way of relating measured contamination
levels resulting from the use of seed treatments to a harmful or fatal outcome and will be useful in considering measured
nectar and pollen concentrations below. This is no different than the USEPA approach of providing “risk quotients,” a ratio of
measured exposure to toxicity, but will allow us to unpack this ratio and look specifically at the assumptions the USEPA uses
to derive expected nectar and pollen concentrations.
Using the consumption rate of 292 mg nectar per foraging bee/day provided by the USEPA assessment17 allows the calculation
of a “limit” concentration considered safe to honey bees if the bee is going to forage on the treated crop for 10 days:
0.00036 ug a.i./bee = “limit” nectar concentration (ug a.i./mg nectar) X 292 mg nectar/bee
Limit nectar concentration (ug a.i./mg nectar) = 0.00036 ug/bee / 292 mg nectar/bee
Therefore, this limit nectar concentration would be equal to 0.00000123 ug a.i./mg nectar or 1.23 ng a.i./g nectar. A review
of expected nectar residue concentration levels in treated crops and surrounding habitat following the use of seed treatments
(below) confirms the high lethal risk to individual foraging bees—whether Apis or non-Apis.
In addition, many papers have examined neonics’ sub-lethal effects on bee learning, reproduction, immune function etc. (see
partial review in Hopwood et al. 2016). These sub-lethal endpoints are critical for assessing risk to wild bee species, whether
solitary bees or even colonial ones given their smaller size and lack of “buffering,” which a honey bee colony with tens of
thousands of individuals enjoys. The 1.23 ng a.i./g nectar value computed above only refers to lethal effects on half of the
population and is therefore not fully protective. The USEPA higher-tiered risk assessments for hive effects do not apply to other colonial wild bee
The USEPA assessment proposes that drone honey bees represent a worst case with respect to pollen ingestion by bees and
equates this value with consumption by nurse bees. This value is given as 3.6 mg/day.18 However, in their guidance document
(EFSA 2013) European regulators point out that pollen ingestion ranges between 26.6 and 30.3 mg/bee/day for bumble bees;
and 10.2 mg/bee/day for solitary bees—in other words, roughly 2.5 to 8 times more than USEPA’s “worst case. Clearly, non-
Apis bees may consume much more pollen than honey bees, making the USEPA assessment an almost certain underestimate
if applied to wild colony-forming bees. Even for honey bees, other authors have used pollen ingestion values in their risk
assessments that are much higher than that proposed in the USEPA assessment (e.g., 9.5 mg/bee/day in Stoner and Eitzer
If we were to use the midpoint value for bumble bee ingestion (28.45 mg/bee/day), and ignoring that bumble bees appear to
be more sensitive than honey bees to clothianidin or thiamethoxam by the oral route, the calculation equivalent to the above
for nectar returns a “limit” pollen value of 12.7 ng a.i./g pollen, clearly much lower than the 97.5 - 372 ng/g NOAEC pollen
values derived from honey bee colony studies (Table 14).
The USEPA assessment does concede (2020d; section that several studies in the open literature indicate risk to bumble
bee colonies at concentrations of imidacloprid much lower than those known to affect honey bee hives. Indeed, at least four
studies reviewed by USEPA documented significant colony effects at exposure levels of 10 ng/g, and one study documented
17 E.g., Table 4.12 in USEPA 2020d. Although we will accept EPAs value for now, it should be noted that this value is not a worst-case scenario. According to the compilation of
EFSA (2012, 2013), sugar requirements in a forager can be as high as 128 mg sugar/day. Given nectar concentrations that range between 15% and 30%, this means that nectar
ingestion could be as high as 427-853 mg nectar/day. Note that the USEPA in their guidance (EPA/PMRA/CaDPR 2014) use the “least protective” value of 30% sugar concen-
tration in nectar.
18 e USEPA assessment compares colony eects to a combined nectar/pollen exposure calculated as the sum of the nectar concentration plus the pollen concentration divided
by 20 given their estimation that honey bees consume 20X more nectar than pollen.
effects at exposure levels as low as 0.7-1.4 ng/g. These exposure levels are much lower than the 19 - 23 ng/g colony effect
levels derived for honey bees. Similarly, USEPA (2020e; pp. 97-98) reviewed studies documenting effects of clothianidin and
thiamethoxam on bumble bee colonies, including Sandrock et al. (2014). That study found that syrup concentrations of 0.45
ng/g of clothianidin or 2.87 ng/g of thiamethoxam significantly decreased reproduction (50% decline) and modified sex ratio
in the red mason bee (Osmia bicornis). Although the entire experiment ran for 35 days, the treated colonies started diverging
from control approximately 6-7 days after the beginning of dosing—less than the 10-day period used in the chronic toxicity
assessments. These effect levels are 7-42 times below the 19 ng/g concentration used by USEPA (2020e) (see Table 15).19
Despite the highly congruent nature of these findings, USEPA dismissed all these studies as unsuitable for “quantitative use in
the risk assessment,” in large part because none of the authors analytically verified the concentration of their dosing solutions
(USEPA 2020d; Table 5-13). This effectively ensured that only industry studies on honey bees were used in the final risk
assessment. Although analytical confirmation of residues in the non-industry studies would have been desirable, it is difficult
to believe that all of the researchers similarly erred and grossly miscalculated dosing levels. Verification of dosing levels is
especially critical when there is the possibility that the active ingredient will degrade over time in the dosing solutions. Here,
any degradation would mean that effect levels would effectively be lower than those reported. Given this, the number of colony
studies and their congruent results, and that the effect levels are in line with chronic toxicity estimates from pollen ingestion as
shown above, to have excluded the entire corpus of independent effect studies was not justified and resulted in USEPA almost
certainly underestimating colony risk to bumble bees.
In addition, this exclusion of independent studies must be weighed against a possible bias in studies sponsored by pesticide
registrants for the purposes of gaining regulatory approval of pesticides (e.g., Boone et al. 2014; Bero et al. 2016; Mie et al
2018; Sheppard et al 2020).
4.4.2. The USEPA assessment underestimates nectar and pollen contamination that results from the
use of seed treatments
The USEPA assessment estimates the expected nectar and pollen neonic residues in certain crops planted with neonic-treated
seed, but the scope of these estimates is limited and underestimates the likely real world residue levels.
In the USEPA assessment, residue values varied considerably between the few crops that were studied. Unfortunately, residue
levels have not been established for the vast majority of crops where the products are registered. This is an unacceptable
situation from a regulatory perspective. Submission of residue data should have been made a condition of registration.
Where information for a specific crop is lacking, USEPA recommends using the combined individual study data for all crops.
For pollen, the sheer number of studies clearly gives a strong weighting to corn, which, incidentally, has the lowest pollen
residue values of all crops studied to date. Soybean and cotton represent most of the data for nectar. Of the few crops studied,
canola (oilseed rape) had the highest pollen and nectar residue values once normalized by application rate per seed (Table 15;
Appendix 6), but these values are overshadowed by the other field crop data. A more precautionary (and I believe appropriate)
approach to that proposed in the USEPA assessment would have been to take the crop with the higher pollen and nectar
residues as a surrogate for those crops without residue information—or at least give equal weighting to the different crops
studied rather than treating each study as an independent sample, which clearly they are not. Indeed, using canola as the
surrogate for crops without residue information would still likely fail to capture the greatest risk. Given that only 4 crops have
been studied (3 for pollen), it is likely that several crops for which neonics are registered will indeed show higher residues in
nectar and pollen than the current maxima generated in canola.
In an earlier compilation, the European Food Safety Authority (EFSA 2013) had already arrived at higher values (Table
16), likely due to consideration of more oilseed rape studies as well as a few sunflower studies.20 However, that compilation
appeared to combine all values together without regard to individual study means or maximum values as done in the USEPA
assessment and is therefore not as robust.
19 Although the NOAEC for thiamethoxam is 44 ng/g (Table 15), USEPA correctly uses the lower value of 19 ng/g for clothianidin because thiamethoxam breaks down to
clothianidin within crop plants.
20 It is also clear that some studies showing higher corn pollen levels were omitted from the USEPA assessment. For example, Bonmatin (2005) recorded maximum corn pollen
levels of 18 ng a.i./g with seed treated at 77% of the currently allowable treatment.
Table 16. Summary of expected pollen and nectar concentrations in crops planted with neonic-treated seeds.
Data for the USEPA 2020 assessment is shown in Appendix 6.
90th percentile concentration (ng a.i./g) normalized to 0.1 mg a.i./seed
USEPA 2020
all crop study
USEPA 2020
all crop study
EFSA 2013
individual values
USEPA 2020
canola only
USEPA 2020
canola only
Pollen 1.8 3.2 4.2 32.7 43.3
Nectar 4.5 7.6 7.7 8.0 11.3
USEPA’s overreliance on corn pollen values and soybean/cotton nectar values is clearly consequential and results in risks
from seed treatments often being downplayed. Table 17 compares the USEPA-estimated pollen and nectar concentrations
with those derived from the precautionary assumption that any given crop might be more canola-like” than “corn-like” or
“soybean-like.” The newly estimated pollen levels alone clearly place bumble bee individuals and colonies at extreme risk in a
number of crops.21
As shown in Table 17, the precautionary approach of taking the crop with the higher pollen and nectar residues as a surrogate
for those crops without residue information would show the risk to be more than an order of magnitude higher for pollen
ingestion; about double for nectar ingestion.
Table 17. Comparison of averagea nectar and pollen concentration estimates between the USEPA assessment approach
(i.e., taking the average of all studies heavily weighted to a few crops) versus an alternative approach that supposes
crops will develop flower residues more in line with canola.
per seed (mg)
mean pollen
based on
data for all
crops heavily
weighted to
corn (ng/g)
mean pollen
if plant is
more “canola-
like” (ng/g)
mean nectar
based on
data for all
crops heavily
weighted to
soybean and
cotton (ng/g)
mean nectar
if plant is
more “canola-
like” (ng/g)
cucurbits 0.750 thiamethoxam 13.5 245 33.75 60
carrot 0.120 clothianidin 2.16 39.2 5.4 9.6
safflower 0.50 imidacloprid 9.0 163 22.5 40
sugar beet 0.720 clothianidin 12.96 235 32.4 57.6
a e mean value given is the 90th higher percentile of means obtained from each study. is is the value recommended to test against a chronic (10-day) exposure for
foraging bees.
In the USEPA assessment, any risk quotient which showed a possible effect for any given crop led to a Tier 2 assessment where
honey bee colony health and survival became the endpoint. This resulted in all risks to bees from seed treatments being
dismissed or considered to be low, often being above an individual NOAEL, but below the defined colony LOAEL. While this
method already likely underestimates risks to solitary bee species, it also is clearly questionable for colony forming species,
given the weakness of USEPA’s underlying assumptions about potential exposure levels.
It should be noted also that the dismissal of risk can be at odds with the available evidence. For example, soybeans were said
to represent no risk when applied as a foliar spray and carry the “weakest evidence of risk” as a seed treatment—yet USEPA
acknowledges three possible or probable bee kill incidents reported in that crop (USEPA 2020d; Table 6-32).
21 Compare estimated pollen levels of 100 or 200+ ng/g in Table 17 with the 10-day limit lethal concentration of 12.7 ng/g calculated above or the many excluded bumblebee
studies showing colony eects at 10 ng/g.
4.4.3. The USEPA assessment ignores critical exposure routes
The USEPA assessment fundamentally considers only two routes of exposure: Direct contact with insecticide droplets at the
time of application (foliar application) as well as the consumption of nectar and pollen containing neonic residues in target
crop plants. This greatly underestimates the actual neonic exposures bees and other pollinators are likely to face from many
exposure routes under real world conditions.
Pollinator assessment schemes (e.g., EPA/PMRA/CaDPR 2014, EFSA 2013) have discussed the need to consider other routes
of exposure such as neonic dust produced by the planting of neonic-treated seeds, plant guttation fluids, and surface water
(including puddles potentially used as drinking water—particularly relevant in California’s predominantly arid and irrigation-
dependent agricultural environment). Other possible routes include honey dew, soil (for ground-nesting bees) and leaves
(leaf-cutting bees). Several of those routes of exposure have already been built into risk calculators for bees (e.g., Mineau
2014). Some of the literature review carried out in the context of this indicator is reproduced here. Dust as a route of exposure
Dust production from the abraded seed treatments during planting is known to be an important source of neonic exposure.
Many kills of honey bee colonies have been reported from this route of exposure in Europe as well as North America.
However, the USEPA assessment only acknowledges, but does not include, this route of exposure in its calculation of risk.
Other researchers have attempted to quantify this risk. For example, Tapparo et al. (2012) conducted experiments where
individual bees trying to reach a food source were captured after flying over a corn field during planting (the entire test
running for 1h). These authors measured amounts of 0.078-1.240 µg/bee (N=5, mean=0.570 µg/bee) for clothianidin at
1.25 mg a.i./seed and 0.128-0.302 µg/bee (N=4, mean=0.189 µg/bee) for thiamethoxam at 1 mg a.i. /seed. The same authors
reported maximum concentrations of 3.65 µg/bee (approximately 36.5 ppm) obtained in previous work with imidacloprid-
treated seed. After a few hours and normal activities in the hive, residues had dropped by an order of magnitude; the bulk of
the insecticide-laden dust on bee surfaces was thought to have been taken back to the hive but dislodged though normal hive
activities. With potentially thousands of foragers returning to a hive, large quantities of insecticide can thus be transferred
efficiently to the hive environment.
Krupke et al. (2012) analyzed samples of honey bees from kills associated with the planting of treated maize seed. Clothianidin
levels in dead and dying bees were much lower than in the Tapparo et al. (2012) study, ranging from 0.0038 to 0.013 ppm. Bees
die from a number of causes not associated with insecticide exposure; Tapparo et al. (2012) also mentioned that dead bees
sampled in the hive after sowing showed no exposure or at least had residue levels that were below detection limits.
From a “residue per unit dose” point of view, it appears that seeding results in higher contamination of insects than an
equivalent spray application (Mineau and Callaghan 2018), but due to the lower per acre rates of application for seed
treatments, neonic concentrations available to pollinators are still lower than following a typical foliar application. The USEPA
assessment relies on a single study by Koch and Weisser (1997) to estimate the amount of residues bees would accumulate
while flying through a cloud of spray droplets. Yet, they ignore very similar data produced by Tapparo et al. (2012) and others
for bees flying through dust generated at seeding.
Mitigation strategies for minimizing the impact of abraded seed dust are not fully implemented or fully effective. In North
America, it is customary for farmers to use talc or graphite as lubricants in their seeding machinery (e.g., Krupke et al. 2012).
Canada has banned the use of these seed lubricants on the grounds that they may cause more abrasion and loss of pesticide
(Health Canada/OMAFRA 2014). Since 2014, any lubricant used in Canada is mandated to be a patented lubricant marketed
by Bayer, which, according to the manufacturer, reduces the amount of clothianidin contamination at seeding (Bayer 2020).
However, the use of a lubricant is not mandatory. Moreover, it has been shown that the efficacy of lubricants at reducing
insecticide dust has probably been exaggerated (Schaafsma et al. 2017) because of deficiencies in the test protocols, namely the
fact that soil dust in the incoming airstream greatly increases the abrasion of the pesticide from the seeds.
Another seed dust mitigation strategy consists of using deflectors which redirects the flow of air from the seeder towards the
ground. This reduces, but does not eliminate, dust drift (see below). In a ground-breaking Canadian study (Tsvetkov et al.
2017) following Canada’s regulations mandating low-fluency agents, in-hive contamination levels were still sufficiently high to
negatively impact colony health (see Section 3.3).
Dust produced at seeding has given rise to visible mortality. To date, seed treatment and soil uses (not foliar uses) comprise
the bulk of imidacloprid bee-mortality incidents reported to USEPA (USEPA 2020f). Yet, both the USEPA and CaDPR
maintain that they are working with different stakeholders to identify best management practices (of the type already found to
be inadequate in Canada) and to promote some sort of as-yet unidentified technological fix to the problem. Schaafsma et al.
(2017) indicated that several corrective actions would be needed to adequately reduce abraded seed exhaust from air seeders;
both intake and exhaust air supplies from air seeders need to be scrubbed and seed polymer coatings need to be improved. At
this point in time, neither USEPA nor CaDPR appear inclined to adopt enforceable regulations to reduce seed dust, and the
cost of technology-based solutions will likely discourage widespread adoption anytime in the near future. While beekeepers
are now aware of this risk and move their honey bee hives to avoid seeding time, this does nothing to protect wild bees from
the impact.
In Europe, risk assessments take dust production at seeding into account. As early as 2013, EFSA had proposed deposition
rates associated with various types of application. EFSA also examined the efficacy of some of the “best management
practices” likely to be proposed eventually by USEPA—e.g., deflectors that direct the dust towards the ground to reduce
exposure to flying insects (see Table 18). For example, without deflectors, the amount of dust drift in field margins during
corn seeding is deemed to be 17% of the applied field rate, very much of similar magnitude as drift documented from high
drift foliar application scenarios, such as fruit tree spray applications. This estimate, however, is likely a worst-case scenario
recommended for risk assessment purposes; the exact rate is highly variable and dependent on a multitude of factors
starting with the seeding equipment (Xue et al. 2015). Regardless, dust production at seeding can be incorporated into an
assessment of risks to bees—albeit with some assumptions as to realistically possible and enforceable improvements in seeding
Table 18. Default deposition percentages for dust drift into field margins to be used for the different combinations of
application technique and types of plants. From EFSA (2013; Appendix H).
Application type Crop type
For purpose of measuring
concentrations in
nectar and pollen
(% of application rate)
For purpose of contact
exposure assessment
(% of application rate)
Spray applications
(spray drift)
Field crops 0.92 2.8
Early fruit 9.7 29.2
Late fruit 5.2 15.7
Early grapevine 0.9 2.7
Late grapevine 2.7 8
Hops 6.4 19.3
Seed treatments
(dust drift)
Maize with deflector 0.56 1.7
Maize without deflector 5.6 17
Oil seed rape with deflector 0.22 0.66
Oil seed rape without
deflector 2.2 6.6
Cereals with deflector 0.33 0.99
Cereals without deflector 3.3 9.9
Sugar beets with deflector 0.001 0.003
Sugar beets without
deflector 0.01 0.03
Granule applications
(dust drift) All crops 3.2 9.6
There is a clear parallel between EPA, PMRA, and CaDPR’s collective failure to adequately assess the risk from seed treatment
uses of neonics and their decision to exempt pesticide-treated seeds from tracking or regulation.
NEONIC SEED TREATMENTS IN CALIFORNIA | 35 Surface water as a route of exposure
The USEPA assessment determined a priori that risk from surface water ingestion was a minor route of exposure for bees (e.g.,
Fig. 2-3 in EPA 2020d), and therefore excluded exposures to neonics from water contact or consumption in its calculation of
risk. Yet, in the context of California’s arid climate where irrigation is often critical to crops, irrigation water and puddles or
other available surface waters are likely critical sources of water for many insect species, including non-Apis bees.
Water needs in bees are expected to be quite variable and thought to be dependent on temperature and local nectar yields.
Most of the research on this subject has been carried out in the honey bee. A low availability of nectar means that water
needs to be obtained from extraneous sources rather than from nectar alone (Kühnholtz and Seeley 1997). This suggests
that extraneous water needs might be higher in intensively farmed landscapes with low nectar yields dominated by corn
or other field crops. Regardless, water needs in spring and early summer are typically large, in part to dilute winter stores
(Butler 1940). At one of their study sites, Kühnholtz and Seeley (1997) noted that the bees favored the muddy wet ground on
the edge of a pond for water collecting. Mineau and Kegley (2014) reported on the observation that bees appeared to prefer
wet muddy ground to a nearby pond. It has been known for a long time (e.g., Butler 1940) that bees are often attracted to
“unsanitary” sources of water, such as rainwater gutters choked with organic debris, sewage effluents, or puddles on top of cow
dung in preference to clean water supplies provided for their use. Butler (1940) was able to confirm that bees preferred some
concentrations of sodium and ammonium chloride to distilled water. However, dilute organic solutions (leaf debris, manure,
and urine) proved more popular still. In the context of an agricultural field, this raises interesting questions. For example,
the attractiveness of water puddles may vary depending on the use of fertilizers (both natural or synthetic) and possibly even
some pesticides (especially dissociated ionic compounds). Finally, Visscher at al. (1996) reviewed older evidence that water-
collecting bees took heavier loads of water when the water was warm; any source of water in fields is likely to heat up faster
than deeper bodies of water when exposed to the sun. These authors calculated that a water collecting bee is restricted to
obtaining water within a 2.1 km radius of the hive based on energetics—compared to the 13.5 km that has been observed for
nectar foragers.
Taken as a whole, these finding suggest that bees will seek out puddles in and around farm fields as a water source, particularly
in the spring planting time. Considerable neonic levels have been detected in these kinds of puddles—likely due to the use
of neonic-treated seeds. Samson-Robert et al. (2014) measured the concentration of pesticides in rain puddles at seeding
(while planting was still in progress) and one month after seeding in corn. The puddles were large ones—described as 1.5-3
sq. m in size and between 4-6 cm in depth. No field spiking (adding a known quantity of pesticide to the samples to ensure
no breakdown or loss in transit) was carried out, so reported values should be considered minimum values. Based on two
years of sampling, all water samples taken from corn fields contained residues of either clothianidin or thiamethoxam; 83%
of samples contained both. Several other pesticides were also detected, but, in samples taken one month after seeding, only
clothianidin, thiamethoxam and the fungicide azoxystrobin were still found at levels exceeding the level of quantification (1
ppb). Levels were higher immediately after seeding, suggesting that dust production during seeding might be an important
pathway by which puddles became contaminated. For clothianidin, mean and maximum concentrations were 4.6 and 56 ng/g;
for thiamethoxam, 7.7 and 63 ng/g. These values are in the range of expected nectar residues documented above. Clearly any
use of in-field water by bees will add substantially to their overall exposure.
EFSA (2013) in their pollinator guidance, recommended using a water consumption figure of 11.4 µL/day per foraging bee
or 111 µL/day per larva, but did not provide further justification for those figures other than to mention they were at the
high end of values obtained from the literature. The USEPA/PMRA/CaDPR guidance (2014) looked at two estimates of water
consumption rates in honey bees. One of those estimates (450-1800 µL/day) was based on direct observations and calculations
from water forager bees. References were supplied to show that between 30-60 µL were collected per foraging trip (e.g., see the
work of Visscher et al. 1996) and that 30% of all water collected was consumed by the bee. However, because these estimates
relate to water foragers and not to other worker bees, and because the estimates work out to a very high (5-20X) turnover of
body water, USEPA favored another estimate, this one based on water flux in a similarly-sized species, the brown paper wasp.
Indeed, their analysis concluded that, depending on conditions and food supply, bee food (i.e., nectar, honey) represents
between 7—>100% of daily water needs. USEPA thus arrived at a maximum water consumption estimate of 47 µL/day, which
the agency recommended for risk assessment purposes; yet, 8 years later, USEPA still does not factor water intake into its
calculation of risk.
The possibility of exposure through water is made more complicated by the fact that pesticides can be absorbed from the
bee’s foregut; i.e., by water foragers bringing water back to the hive, rather than drinking it for their own water needs (Conner
et al. 1978). Absorption was found to be highest at low sucrose concentrations—i.e., the situation in a water forager vs. the
usual test situation in oral toxicity tests. It is known also that worker bees do collect water to cool the hive. This may not be
equivalent to drinking exposure, but does undoubtedly lead to difficult-to-measure exposure.
NEONIC SEED TREATMENTS IN CALIFORNIA | 36 Guttation water as a route of exposure
The USEPA assessment (e.g., Fig. 2-3 in EPA 2020d) also determined a priori that risk from consumption of neonic-
contaminated guttation fluids (the excretion of excess water or nutrients through small openings on a plant’s leaves or stems)
was minor, likewise excluding it from the risk calculation. However, research on the high concentration of neonics in the
guttation fluid of treated crop plants indicates that these fluids, if consumed, would constitute a highly significant exposure
Several researchers have documented concentrations of various neonics in guttation water following their use as seed
treatments in corn (Table 19). They reported that, on corn plants, experimenters were able to reliably and easily collect
guttation droplets for at least three weeks after seeding under field conditions. Unlike what had been suggested in the
literature, and assumed by regulatory authorities regarding this exposure route, they found that the phenomenon was not
restricted to situations of high soil moisture and high humidity; moreover, droplets tended to pool in the leaf whorl of the
developing plant. Only evaporation reduced the availability of droplets; however, the researchers proposed that concentrations
could increase over time following repeated drying and droplet formation cycles.
Table 19. Measured concentrations of neonicotinoids and fipronil in guttation water from seed-treated corn.
Rate of
a.i. per
seed (mg)
Mean (SE)
or range
(days 1-6 after
reported maxima
normalized to
0.1 mg/seed (ng
a.i./g) Reference
imidacloprid 0.5 (field) 47 (9.96) >200 9,400 Girolami et al. 2009
imidacloprid 0.5 (pots) 82.8 (14.07) 16,600 Girolami et al. 2009
imidacloprid 1.25 103-346 (leaf tip) 346 8,200-27,700 Tapparo et al. 2011
imidacloprid 1.25 8.2-120 (whorl) 120 660-9,600 Tapparo et al. 2011
clothianidin 1.25 23.3 (4.2) >100 1,900 Girolami et al. 2009
clothianidin 1.25 76-102 (leaf tip) 102 6,100-8,200 Tapparo et al. 2011
clothianidin 1.25 7.3-47 (whorl) 47 580-3,800 Tapparo et al. 2011
clothianidin 1.25 7.5 - 8 600-640 Reetz et al. 2011
thiamethoxam 1 12 (3.3) >100 1,200 Girolami et al. 2009
thiamethoxam 1 16-41 (leaf tip) 41 1,600-4,100 Tapparo et al. 2011
thiamethoxam 1 2.9-26 26 290-2,600 Tapparo et al. 2011
Comparison between these values and those predicted in nectar and pollen shows that the neonic concentrations in guttation
droplets are typically two to three orders of magnitude higher. Accordingly, consumption of even a small amount of this water
would dramatically change the risk picture for pollinators.
In its proposed problem formulation for pollinator risk assessments, USEPA (2012) downplays pollinator exposure risks from
drinking neonic-contaminated water because: (1) some of those sources such as dew, puddles, or guttation droplets are not
always present and ephemeral when present; and (2) the majority of foraging bees are expected to obtain most of their water
needs through nectar. However, USEPA acknowledges that, if water was indeed to be obtained through puddles or guttation
fluids, these routes of exposure would dwarf other routes of exposure such as direct sprays or dietary exposure through
nectar or pollen. Regardless, USEPA/PMRA/CDPR (2014), in their final guidance document, opted not to include risks from
exposure to neonics from drinking water, including guttation fluids.
In contrast, EFSA (2013) recommended that guttation water be included in the first tier of assessment, but that there also be
an assessment of the likelihood of guttation droplet formation based on location conditions and calendar date.
I concur with EFSA (2013) and Blacquière et al. (2012) that prudence requires that drinking water routes of exposure be
considered, at least until more information is obtained on their real world importance. This should be done not just for honey
bees, but other bee species as well. The USEPA and CaDPR risk assessments are likely dramatically underestimating true risks
to pollinators if potential neonic exposures from water sources, like guttation fluid, are excluded.
4.5. Risks and impacts to pollinators from neonic-treated seeds are not
restricted to the crop area
Despite their recent analysis of risk for pollinators, both USEPA and CaDPR fail to accurately account for risks from neonic-
treated seeds and, arguably, underestimate risks from neonics more generally. Failure to characterize the risk from seed
treatments is a huge oversight given the potentially large area where treated seeds are planted, notwithstanding that higher risk
may result from other application methods—notably, foliar and ground sprays.
Initially, the USEPA and CaDPR assessments incorrectly assume that neonic active ingredients on treated seeds remain
restricted to the field area. As such, the assessment excludes crops that are not attractive to pollinators or that are harvested
before flowering; in doing so, the assessments fail to consider the movement of residues off site—whether through dust,
runoff, or windblown soil—into the pollen and nectar of adjacent wildflowers. This movement, however, is well documented.
For example, Stewart et al. (2014) looked at seed treatments in corn, cotton, and soybean. They found evidence of
contamination of wildflowers situated 20m on average from seeded fields. Residues were detected in a quarter of wildflower
samples and, when detected, averaged 10 ng/g. This is about 10X the limit concentration calculated above as causing mortality
in honey bees after 10 days of feeding. Rundlöf et al. (2015) similarly documented clothianidin in pollen and nectar in wild
plants in the edges of oilseed rape sowed with a seed treatment containing clothianidin. Botias et al. (2016) documented
significant contamination of wildlflowers in field edges near oilseed rape fields, which had an average residue level of 14.8 ng/g
of thiamethoxam, while the canola flowers in the treated field had an average residue level of 3.26 ng/g. Higher levels still were
recorded on the foliage of wild plants in the margins; and, similarly, at greater levels than in crop foliage (Botias et al. 2016).
Long and Krupke (2016) likewise found that pollen originating from a large number of non-crop plants in field borders was
contaminated with a multitude of pesticides, not all from the adjoining crop. Finally, Bredeson and Lundgren (2019) found
high residues in cover crops inter-seeded with corn; which is perhaps ironic given that the practice is intended in part to
promote beneficial insect communities.
4.6. Harmful impacts from neonic-treated seeds have already been
Importantly, the risks and harms posed by the use of neonic-treated seeds are not purely theoretical. Studies have shown
adverse impacts at the levels of neonic contamination documented following seed treatment uses.
Tsvetkov et al. (2017)—published in the prestigious journal “Science”—showed that bee food stores were heavily
contaminated with clothianidin or thiamethoxam when hives were placed near corn fields having used those seed treatments.
They then carried out a controlled exposure study that showed that measured levels of contamination were sufficient to
negatively affect colony health. Indeed, their chosen levels of dosing (4.9 ng/g initially, declining to 2.0 ng/g of clothianidin on
pollen patties) were below the measurements of actual hive exposures. These levels are also a fraction of the levels considered
to be safe based on the industry studies used in the CaDPR and USEPA assessment (16 ng/g for a total nectar and pollen
blended exposure; 372 ng/g for pollen alone). What also makes the Tsetkov et al. study particularly compelling is that the
statistical analysis was performed under blind conditions, and that all of the raw data has been made available from a data
depository site. Also, the study took place well after the mandated requirement for better fluency agents in Canada (Health
Canada/OMAFRA 2004). It therefore highlights a situation that is already much more protective against neonic exposures to
pollinators from abraded seed dust than the current status quo in the United States. In agreement with the work of Botias et
al. (2015) and Long and Krupke (2016), the bulk of pollen stores were from non-crop plants. There is no indication that this
study was considered in the 2020 USEPA assessment, despite its 2017 publication date. Several industry-submitted studies,
however, with much higher effect levels, were considered.
Further, while it is typically difficult to tease out the effect of a specific stressor in ecological field studies, this has been done in
several cases with neonic seed treatments.
Gilburn et al. (2015) noted a strong association between an increase in the area of oilseed rape (canola) treated with neonic
seed treatments and significant decreases in population counts for 17 species of butterflies in the United Kingdom (UK). Like
the analysis of Hallman et al. (2014) for insectivorous birds, Gilburn et al. were able to show that, prior to the introduction
of neonics (1985-1998), butterfly numbers were actually increasing in those same areas making it less likely that butterfly
numbers were simply reflective of intensive cropping.
Also in the UK, Woodcock et al. (2016) found that distribution data for 62 wild bee species showed an increase in the
probability of local population extinction rates in areas of oilseed rape seeded with neonics. The most affected species were
those foraging in the crop (it is useful at this point to recall that the USEPA assessment considers risk from canola to be low).
Importantly, Woodcock et al. (2016) also showed that the use of foliar insecticides did not appear to have an effect on bee
extinction probability, however, several best management practices implemented in English farming systems specifically to
protect pollinators may have helped reduce the impact of sprays.
In California, Forister et al. (2016) , working on four different long-term monitoring sites, were able to associate butterfly
declines to neonic use measured on a county basis. The authors were able to remove the effects of other insecticides as well to
isolate the effect of neonic use; however, their pesticide data came from the PUR which means that seed treatments were not
included in their pesticide use estimates.
Rundlöf et al. 2015 found that oilseed rape seed treatments containing both clothianidin and the pyrethroid beta cyfluthrin
affected both bumble bees (Bombus terrestris) and the solitary bee Osmia bicornis in field borders of Swedish farms. Because
beta-cyfluthrin in not systemic and tighly bound to soil, the authors surmised that the impacts they were seeing were
primarily from the clothianidin component.
A recent study (Main et al. 2020) looked at native bee diversity in and around corn or soybean in rotation treated with
imidacloprid or clothianidin. Although they found less contamination of flowers in field edges than Botias et al. (2015),
they measured lower diversity of wild bees associated with treated fields. Even their untreated fields showed the presence of
neonicotinoid residues, although lower and not as frequent. The presence of a diverse wildflower community proved to be the
most important factor controlling wild bee abundance—but not diversity.
An industry-funded study in Germany (Peters et al. 2016) failed to find an impact from the same clothianidin & beta-
cyfluthrin product examined by Rundlöf et al. (2015), and reported levels of hive contamination were lower. Another
industry-funded study (Sterk et al. 2016) similarly did not find any effects from this same seed treatment on bumble bees.
Again, the authors commented on the fact that levels of contamination in their study were lower than in non-industry studies.
This suggests that there were more alternative resources available for the bees in the German industry studies. Ultimately, it
appears that whether or not bees have access to a large quantity of clean floral resources well away from neonic-treated fields
or whether they are more restricted to feeding within immediate field borders greatly affects risk. The greater risk associated
with floral resources in or near treated fields, however, appears driven by the use of neonic-treated seeds as well as foliar and
soil applications.
4.7. Conclusions of the pollinator assessment
In summary, the CaDPR and USEPA assessments greatly underestimate the risks to pollinators from the use of
nitroguanadine-neonic-treated seeds. In particular, the assessments: (1) underestimate risks to wild bee species and other
pollinators by relying on honey bee colony survival as a proxy for overall pollinator health; (2) underestimate nectar and
pollen contamination levels following the use of neonic-treated seeds by assuming that the majority of crop species will
have residue values at the low end of the measured spectrum; (3) ignore risks of dust from neonic-treated seeds at planting,
despite ample evidence that this route of exposure is highly relevant; (4) ignore exposures of bees and other pollinators to
neonic contaminated water—including, guttation fluid and puddles in or near fields sown with neonic-treated seeds—despite
existing field estimates that show that these routes of exposure can completely dwarf the routes that have been formally
assessed; and (5) ignore risks from neonic uses on crops deemed unattractive to honey bees, despite evidence that neonic
residues migrate into adjoining areas, including adjacent wildflowers that can exceed levels in the field proper; (6) exclude
available peer-reviewed literature from quantitative risk assessment in favor of industry studies; and (7) ignore the growing
amount of field data which now links the use of neonic-treated seeds to pollinator failure on a landscape scale. The USEPA
and CaDPR therefore fail to appreciate and acknowledge the likely considerable and damaging effect that neonic-treated seeds
are having on California’s pollinator populations.
AG Professional. 2013. Syngenta introduces CruiserMaxx alfalfa seed treatment. February 22, 2013 https://www.agprofessional.
com/article/syngenta-introduces-cruisermaxx-alfalfa-seed-treatment (accessed June 2020).
Alford, A., Krupke, C.H. 2017. Translocation of the neonicotinoid seed treatment clothianidin in maize. PLoS ONE 12,
Allington, A. 2020. When is a pesticide not a pesticide? When it coats a seed. Bloomberg. Environment & Energy Report. Jan.
27 2020.
Arena, M., Sgolastra, F. 2014. A meta-analysis comparing the sensitivity of bees to pesticides. Ecotoxicology. https://doi.
Bayer Corp. 2019. Fluency agent advanced. Sales brochure. 2 pp.
Bero, L., Anglemyer, A., Vesterinen, H., Krauth, D. 2016. The relationship between study sponsorship, risks of bias, and
research outcomes in atrazine exposure studies conducted in non-human animals: Systematic review and meta-analysis.
Environment International 92–93, 597–604.
Blacquière, T., Smagghe, G., Gestel, C.A.M. van, Mommaerts, V. 2012. Neonicotinoids in bees: a review on concentrations,
side-effects and risk assessment. Ecotoxicology 21, 973992.
Bonmatin, J.M., Marchand, P.A., Charvet, R., Moineau, I., Bengsch, E.R., Colin, M.E. 2005. Quantification of Imidacloprid
Uptake in Maize Crops. Journal of Agricultural and Food Chemistry 53, 5336–5341.
Boone, M.D., Bishop, C.A., Boswell, L.A., Brodman, R.D., Burger, J., Davidson, C., Gochfeld, M., Hoverman, J.T., Neuman-Lee,
L.A., Relyea, R.A., Rohr, J.R., Salice, C., Semlitsch, R.D., Sparling, D., Weir, S. 2014. Pesticide Regulation amid the Influence of
Industry. BioScience 64, 917–922.
Botías, C., David, A., Horwood, J., Abdul-Sada, A., Nicholls, E., Hill, E., Goulson, D. 2015. Neonicotinoid Residues in
Wildflowers, a Potential Route of Chronic Exposure for Bees. Environmental Science & Technology 49, 1273112740. https://
Botías, C., David, A., Hill, E.M., Goulson, D. 2016. Contamination of wild plants near neonicotinoid seed-treated crops,
and implications for non-target insects. Science of The Total Environment 566567, 269278.
Bredeson, M.M., Lundgren, J.G. 2019. Neonicotinoid insecticidal seed-treatment on corn contaminates interseeded cover
crops intended as habitat for beneficial insects. Ecotoxicology 28, 222–228.
Budge, G.E., Garthwaite, D., Crowe, A., Boatman, N.D., Delaplane, K.S., Brown, M.A., Thygesen, H.H., Pietravalle, S. 2015.
Evidence for pollinator cost and farming benefits of neonicotinoid seed coatings on oilseed rape. Scientific Reports 5, 12574.
Butler, C.G. 1940. The choice of drinking water by the honeybee. Journal of Experimental Biology 17, 253–261.
CaDPR (California Department of Pesticide Regulation). 2018. Neonicotinoid Risk Determination. J. Troiano et al. 1169 pp.
Pesticide Registration Branch.
CaDPR. 2019. Addendum to the July 2018 California Neonicotinoid Risk Determination. Memorandum from R. Darling, 30
January 2019, 14 pp.
Cavallaro, M.C., Morrissey, C.A., Headley, J.V., Peru, K.M., Liber, K. 2017. Comparative chronic toxicity of imidacloprid,
clothianidin, and thiamethoxam to Chironomus dilutus and estimation of toxic equivalency factors: Chronic neonicotinoid
toxicity and toxic equivalency factors. Environ Toxicol Chem 36, 372–382.
Chen, W., Hertl, P., Chen, S., and Tierney, D. 2002. A pesticide surface water mobility index and its relationship with
concentrations in agricultural drainage watersheds. Environmental Toxicol. Chem., Vol. 21, No. 2, 298–308.
Conner, W.E., Wilkinson, C.F., Morse, R.A. 1978. Penetration of insecticides through the foregut of the honeybee (Apis
mellifera L.). Pesticide Biochemistry and Physiology 9, 131–139.
Douglas, M.R., Tooker, J.F. 2015. Large-scale deployment of seed treatments has driven rapid increase in use of neonicotinoid
insecticides and preemptive pest management in U.S. field crops. Environmental Sci. Technol. 150402080236006. https://doi.
Douglas, M.R., Rohr, J.R., Tooker, J.F. 2015. Neonicotinoid insecticide travels through a soil food chain, disrupting biological
control of non-target pests and decreasing soya bean yield. J. Applied Ecol. 52, 250–260.
Douglas, M.R., Tooker, J.F. 2016. Meta-analysis reveals that seed-applied neonicotinoids and pyrethroids have similar negative
effects on abundance of arthropod natural enemies. PeerJ 4, e2776.
EFSA (European Food Safety Authority). 2013. EFSA Guidance Document on the Risk Assessment of Plant Protection
Products on Bees (Apis Mellifera, Bombus Spp. and Solitary Bees). EFSA Journal 11, no. 7 (2013): 3295.
Forister, M.L., Cousens, B., Harrison, J.G., Anderson, K., Thorne, J.H., Waetjen, D., Nice, C.C., De Parsia, M., Hladik, M.L.,
Meese, R., van Vliet, H., Shapiro, A.M. 2016. Increasing neonicotinoid use and the declining butterfly fauna of lowland
California. Biology Letters 12, 20160475.
Gilburn, A.S., Bunnefeld, N., Wilson, J.M., Botham, M.S., Brereton, T.M., Fox, R., Goulson, D. 2015. Are neonicotinoid
insecticides driving declines of widespread butterflies? PeerJ 3, e1402.
Girolami, V., Mazzon, L., Squartini, A., Mori, N., Marzaro, M., Di bernardo A., Greatti, M., Giorio, C., Tapparo, A. 2009.
Translocation of Neonicotinoid Insecticides From Coated Seeds to Seedling Guttation Drops: A Novel Way of Intoxication for
Bees. Journal of Economic Entomology 102, 1808–1815. doi:10.1603/029.102.0511.
Goh, K.S., Yuzhou, L., and Singhasemanon, N. 2019. Surface Water Protection Program for Pesticide Use in California IN Goh
et al.; Pesticides in Surface Water: Monitoring, Modeling, Risk Assessment, and Management. ACS Symposium Series 1308,
Hallmann, C.A., Foppen, R.P.B., van Turnhout, C.A.M., de Kroon, H., Jongejans, E. 2014. Declines in insectivorous birds are
associated with high neonicotinoid concentrations. Nature.
Health Canada/OMAFRA (Ontario Ministry of Agriculture and Food and Rural Affairs). 2014. Pollinator protection and
responsible use of insecticide-treated seed. 2 pp.
Hitaj, C., Smith, D.J., Code, A., Wechsler, S., Esker, P.D., Douglas, M.R. 2020. Sowing Uncertainty: What We Do and Don’t
Know about the Planting of Pesticide-Treated Seed. BioScience 70, 390–403.
Hladik, M.L., Kolpin, D.W., Kuivila, K.M. 2014. Widespread occurrence of neonicotinoid insecticides in streams in a high corn
and soybean producing region, USA. Environmental Pollution 193, 189–196.
Hokkanen, H.M.T., Menzler-Hokkanen, I., Keva, M. 2017. Long-term yield trends of insect-pollinated crops vary regionally
and are linked to neonicotinoid use, landscape complexity, and availability of pollinators. Arthropod-Plant Interactions 11,
Hopwood, J., Code, A., Vaughan, M., Biddinger, D., Shepherd, M., Hoffman Black, S., Lee-Mäder, E., and Mazzacano, C. 2016.
How neonicotinoids can kill bees. The science behind the role these insecticides play in harming bees. Xerces Society for
Invertebrate Conservation 2nd edition Revised and Expanded, 84 pp.
Hoyle, S. and Code, A. 2016. Neonicotinoids in Californias Surface Waters. A Preliminary Review of Potential Risk to Aquatic
Invertebrates. Xerces Society for Invertebrate Conservation. November 2016. 17 pp.
Jenkins, P.T. 2017. Citizen’s petition to the United States Environmental Protection Agency. 46 pp.
Jeschke, P., Nauen, R., Schindler, M., Elbert, A. 2011. Overview of the Status and Global Strategy for Neonicotinoids. Journal of
Agricultural and Food Chemistry 59, 2897–2908.
Koch, H., Weisser, P. 1997. Exposure of honey bees during pesticide application under field conditions. Apidologie 28, 439–
Krupke, C.H., Hunt, G.J., Eitzer, B.D., Andino, G., Given, K. 2012. Multiple Routes of Pesticide Exposure for Honey Bees
Living Near Agricultural Fields. PLoS ONE 7, e29268. doi:10.1371/journal.pone.0029268.
Krupke, C.H., Alford, A.M., Cullen, E.M., Hodgson, E.W., Knodel, J.J., McCornack, B., Potter, B.D., Spigler, M.I., Tilmon,
K., Welch, K. 2017. Assessing the value and pest management window provided by neonicotinoid seed treatments for
management of soybean aphid (Aphis glycines Matsumura) in the Upper Midwestern United States: Value of neonicotinoid
seed treatments for soybean aphid management. Pest Management Sci. 73, 2184–2193.
Kühnholz, S., Seeley, T.D. 1997. The control of water collection in honey bee colonies. Behavioral ecology and sociobiology 41,
Labrie, G., Gagnon, A.-È., Vanasse, A., Latraverse, A., Tremblay, G. 2020. Impacts of neonicotinoid seed treatments on soil-
dwelling pest populations and agronomic parameters in corn and soybean in Quebec (Canada). PLoS ONE 15, e0229136.
Long, E.Y., Krupke, C.H. 2016. Non-cultivated plants present a season-long route of pesticide exposure for honey bees. Nature
Communications 7, 11629.
Lundh, A., Lexchin, J., Mintzes, B., Schroll, J.B., Bero, L. 2017. Industry sponsorship and research outcome. Cochrane Database
of Systematic Reviews 2017, Issue2. Art.No.:MR000033.
Macfadyen, S., Hardie, D.C., Fagan, L., Stefanova, K., Perry, K.D., DeGraaf, H.E., Holloway, J., Spafford, H., Umina, P.A.
2014. Reducing Insecticide Use in Broad-Acre Grains Production: An Australian Study. PLoS ONE 9, e89119. https://doi.
Main, A.R., Headley, J.V., Peru, K.M., Michel, N.L., Cessna, A.J., Morrissey, C.A. 2014. Widespread use and frequent detection
of neonicotinoid insecticides in wetlands of Canada’s Prairie Pothole Region. Plos One 9, e92821.
Main, A.R., Webb, E.B., Goyne, K.W., Mengel, D. 2020. Reduced species richness of native bees in field margins associated
with neonicotinoid concentrations in non-target soils. Agriculture, Ecosystems & Environment 287, 106693. https://doi.
Maxim, L. and van der Sluijs, J. 2013. Seed-dressing systemic insecticides and honeybees. Chapter 16 IN Late lessons from
early warnings: science, precaution, innovation. European Environment Agency. EEA Report No. 1/2013, 40 pp.
Mie, A., Rudén, C., Grandjean, P. 2018. Safety of Safety Evaluation of Pesticides: developmental neurotoxicity of chlorpyrifos
and chlorpyrifos-methyl. Environ Health 17, 77.
Miles, J.C., Hua, J., Sepulveda, M.S., Krupke, C.H., Hoverman, J.T. 2017. Effects of clothianidin on aquatic communities:
Evaluating the impacts of lethal and sublethal exposure to neonicotinoids. Plos ONE 12(3), e0174171. Https://doi.
Miles, J.C., Hua, J., Sepulveda, M.S., Krupke, C.H., Hoverman, J.T. 2018. Correction: Effects of clothianidin on aquatic
communities: Evaluating the impacts of lethal and sublethal exposure to neonicotinoids. PLoS ONE 13, e0194634. https://doi.
Mineau, P. and C. Palmer. 2013. The impact of the nation’s most widely used insecticides on birds. Report prepared for the
American Bird Conservancy, March 2013. 96 pp.
Mineau, P. 2014. Pesticide Risk Tool (PRT). Acute Pollinator Risk Index. White Paper. October 2014. Prepared for the
Integrated Pest Management Institute. 61 pp.
Mineau P, and Kegley S. 2014. New Science on Neonicotinoids. Pesticide Research Institute. May 28, 2014. https://www.
Mineau, P. and Callaghan, C. 2018. Neonicotinoid insecticides and bats: an assessment of the direct and indirect risks.
Canadian Wildlife Federation. 87 pp.
Mineau, P. 2019a. Impacts of Neonics in New York Water. Their Use and Threats to the State’s Aquatic Ecosystems. Technical
Report prepared for the Natural Resource Defense Council (NRDC). 18 pp.
Mineau, P. 2019b. An Assessment of Neonicotinoid Insecticides with Emphasis on New York: Use, Contamination, Impacts
on Aquatic Systems, and Agronomic Aspects. Technical Report prepared for the Natural Resource Defense Council (NRDC),
Morrissey, C.A., Mineau, P., Devries, J.H., Sanchez-Bayo, F., Liess, M., Cavallaro, M.C., and Liber, K. 2015. Neonicotinoid
contamination of global surface waters and associated risk to aquatic invertebrates: A review. Environ. International 74, 291-
Murray, J. 2015. Neonicotinoid pesticides: Not just a bee problem. City of Santa Barbara, Creeks Division, California
Stormwater Quality Association Conference. 6 pp.
NRC (National Research Council of the National Academies). 2015. Review of California’s risk-assessment process for
pesticides. National Academies Press. 56 pp.
Peters, B., Gao, Z., Zumkier, U. 2016. Large-scale monitoring of effects of clothianidin-dressed oilseed rape seeds on
pollinating insects in Northern Germany: effects on red mason bees (Osmia bicornis). Ecotoxicology 25, 1679–1690. https://
PMRA (Pest Management Regulatory Agency). 2018a. Special Review of Thiamethoxam Risk to Aquatic Invertebrates:
Proposed Decision for Consultation. Proposed Special Review Decision PSRD2018-02. 211 pp.
PMRA. 2018b. Special Review of Clothianidin Risk to Aquatic Invertebrates: Proposed Decision for Consultation. Proposed
Special Review Decision PSRD2018-01. 189 pp.
Raby, M., Nowierski, M., Perlov, D., Zhao, X., Hao, C., Poirier, D.G., Sibley, P.K. 2018a. Acute toxicity of 6 neonicotinoid
insecticides to freshwater invertebrates: Aquatic toxicity of neonicotinoid insecticides. Environ Toxicol Chem 37, 1430–1445.
Raby, M., Zhao, X., Hao, C., Poirier, D.G., Sibley, P.K. 2018b. Chronic toxicity of 6 neonicotinoid insecticides to Chironomus
dilutus and Neocloeon triangulifer: Chronic toxicity of neonicotinoids to aquatic invertebrates. Environ Toxicol Chem 37,
Reed, D.A., T. M. Perring, J. P. Newman, J. A. Bethke, J. N. Kabashima. 2014. Bagrada bug. Pest Notes, University of California,
Publication 74166, 4 pp.
Reetz, J.E., Zühlke, S., Spiteller, M., Wallner, K. 2011. Neonicotinoid insecticides translocated in guttated droplets of seed-
treated maize and wheat: a threat to honeybees? Apidologie 42, 596–606. doi:10.1007/s13592-011-0049-1.
Rundlöf, M., Andersson, G.K.S., Bommarco, R., Fries, I., Hederström, V., Herbertsson, L., Jonsson, O., Klatt, B.K., Pedersen,
T.R., Yourstone, J., Smith, H.G. 2015. Seed coating with a neonicotinoid insecticide negatively affects wild bees. Nature. https://
Samson-Robert, O., Labrie, G., Chagnon, M., Fournier, V. 2014. Neonicotinoid-Contaminated Puddles of Water Represent a
Risk of Intoxication for Honey Bees. PLoS ONE 9, e108443.
Sánchez-Bayo, F., Tennekes, H.A. 2020. Time-Cumulative Toxicity of Neonicotinoids: Experimental Evidence and Implications
for Environmental Risk Assessments. IJERPH 17, 1629.
Sandrock, C., Tanadini, L.G., Pettis, J.S., Biesmeijer, J.C., Potts, S.G., Neumann, P. 2014. Sublethal neonicotinoid insecticide
exposure reduces solitary bee reproductive success: Loss of pollinator fitness. Agricultural and Forest Entomology 16, 119–128.
Schaafsma, A., Limay-Rios, V., Baute, T., Smith, J., Xue, Y. 2015. Neonicotinoid insecticide residues in surface water and soil
associated with commercial maize (corn) fields in Southwestern Ontario. Plos ONE 10(2), e0118139. DOI:10.1371/journal.
Schaafsma, A.W., Limay-Rios, V., Forero, L.G. 2018. The role of field dust in pesticide drift when pesticide-treated maize
seeds are planted with vacuum-type planters: Field dust and seed pesticide drift. Pest. Manag. Sci 74, 323–331. https://doi.
Sheppard, L., McGrew, S., Fenske, R.A. 2020. Flawed analysis of an intentional human dosing study and its impact on
chlorpyrifos risk assessments. Environment International 143, 105905.
Smith, R., A. Baameur, M. Bari, M. Cahn, D. Giraud, E. Natwick, and E. Takele. 2008. Artichoke production in California.
University of California Vegetable Research & Information Center. 6 pp.
Starner, K. and Goh, K.S. 2012. Detections of the neonicotinoid insecticide imidacloprid in surface waters of three agricultural
regions of California, USA, 2010-2011. Bull. Environ. Contam. Toxicol. 88(3), 316-21.
Sterk, G., Peters, B., Gao, Z., Zumkier, U. 2016. Large-scale monitoring of effects of clothianidin-dressed OSR seeds on
pollinating insects in Northern Germany: effects on large earth bumble bees (Bombus terrestris). Ecotoxicology 25, 1666–
Stewart, S.D., Lorenz, G.M., Catchot, A.L., Gore, J., Cook, D., Skinner, J., Mueller, T.C., Johnson, D., Zawislak, J., Barber, J.
2014. Potential Exposure of Pollinators to Neonicotinoid Insecticides from the Use of Insecticide Seed Treatments in the Mid-
Southern U. S. Environmental Science & Technology 140710115101004.
Stoner, K.A., Eitzer, B.D. 2013. Using a Hazard Quotient to Evaluate Pesticide Residues Detected in Pollen Trapped from
Honey Bees (Apis mellifera) in Connecticut. PLoS ONE 8, e77550. doi:10.1371/journal.pone.0077550.
Suchail, S., Guez, D., Belzunces, L.P. 2001. Discrepancy between acute and chronic toxicity induced by imidacloprid and its
metabolites in Apis mellifera. Environmental toxicology and chemistry 20, 2482–2486.
Tapparo, A., Giorio, C., Marzaro, M., Marton, D., Soldà, L., Girolami, V. 2011. Rapid analysis of neonicotinoid insecticides in
guttation drops of corn seedlings obtained from coated seeds. Journal of Environmental Monitoring 13, 1564. https://doi.
Tapparo, A., Marton, D., Giorio, C., Zanella, A., Soldà, L., Marzaro, M., Vivan, L., Girolami, V. 2012. Assessment of the
Environmental Exposure of Honeybees to Particulate Matter Containing Neonicotinoid Insecticides Coming from Corn
Coated Seeds. Environmental Science & Technology 46, 25922599. doi:10.1021/es2035152.
Tsvetkov, N., Samson-Robert, O., Sood, K., Patel, H.S., Malena, D.A., Gajiwala, P.H., Maciukiewicz, P., Fournier, V., Zayed, A.
2017. Chronic exposure to neonicotinoids reduces honey bee health near corn crops. Science 356, 1395–1397.
USEPA (US Environmental Protection Agency). 2010. Acres Planted per Day and Seeding Rates of Crops Grown in the United
States. Memorandum by J. Becker, 10 November 2010. 85 pp.
USEPA. 2012. White Paper in Support of the Proposed Risk Assessment Process for Bees. Submitted to the FIFRA Scientific
Advisory Panel for Review and Comment September 11—14, 2012. Office of Chemical Safety and Pollution Prevention
Office of Pesticide Programs, Environmental Fate and Effects Division, Environmental Protection Agency, Washington
DC; Environmental Assessment Directorate, Pest Management Regulatory Agency, Health Canada, Ottawa, CN; California
Department of Pesticide Regulation.!documentDetail;D=EPA-HQ-OPP-2012-0543-0004.
USEPA. 2016. Preliminary Aquatic Risk Assessment to Support the Registration Review of Imidacloprid. 22 December 2016.
218 pp.
USEPA. 2017. Benefits of Neonicotinoid Insecticide Use in the Pre-Bloom and Bloom Periods of Cotton. 21 Nov. 2017. 33 pp.
USEPA. 2018. Benefits and Impacts of Potential Mitigation for Neonicotinoid Seed Treatments on Small Grains, Vegetables,
and Sugarbeet Crops. 30 August 2018. 19 pp.
USEPA. 2019a. Benefits of Neonicotinoid Use and Impacts of Potential Risk Mitigation in Vegetables, Legumes, Tree Nuts,
Herbs, and Tropical and Subtropical Fruit Crops. T. Harty and T.J. Wyatt. 38 pp.
USEPA. 2019b. Biological and Economic Analysis Division’s (BEAD) Response to Comments on the Preliminary Risk
Assessments and Benefit Assessments for Citrus, Cotton, Soybean Seed Treatment, and Other Crops Not Assessed for
Neonicotinoid Insecticides. B. Gelso et al. 21 pp.
USEPA. 2019c. Usage and Benefits of Neonicotinoid Insecticides in Rice and Response to Comments. K. Tindall and C. Rahtz.
17 pp.
USEPA. 2019d.
risk. Accessed January 2019.
USEPA. 2020a. Updated Screening Level Usage Analysis (SLUA) Report for Clothianidin (044309). Memorandum by L.
Hendrick. 16 January 2020. 6 pp.
USEPA. 2020b. Updated Screening Level Usage Analysis (SLUA) Report for Thiamethoxam (060109). Memorandum by L.
Hendrick. 16 January 2020. 6 pp.
USEPA. 2020c. Comparative analysis of Aquatic Invertebrate Risk Quotients generated for neonicotinoids using Raby et al.
(2018) toxicity data. Memorandum. K. Garber, M. Wagman, K. Sappington, E. Donovan. 7 January 2020. 26 pp.
USEPA. 2020d. Final Bee Risk Assessment to Support the Registration Review of Imidacloprid. Memorandum by K.G.
Sappington et al. 14 January 2020. 327 pp.
USEPA. 2020e. Final Bee Risk Assessment to Support the Registration Review of Clothianidin and Thiamethoxam.
Memorandum by K. Garber et al. 14 January 2020. 229 pp.
USEPA. 2020f. Appendices to the Final Bee Risk Assessment for Clothianidin (PC code 044309) and Thiamethoxam (PC code
060109). 246 pp.
USEPA. 2020g. Attachment 1 to the Neonicotinoid Final Bee Risk Assessments. Tier II Method for Assessing Combined Nectar
and Pollen Exposure to Honey Bee Colonies. 16 pp.
USEPA. 2020h. Attachment 2 to the neonicotinoid final bee risk assessments. Residue Bridging Analysis of Foliar and Soil
Agricultural Uses of Neonicotinoids. 226 pp.
USEPA. 2020i. Attachment 3 to the Neonicotinoid Final Bee Risk Assessments. Residue Bridging Analysis for Foliar and Soil
Non-Agricultural Uses of Neonicotinoids. 30 pp.
USEPA. 2020j. Attachment 4 to the Neonicotinoid Final Bee Risk Assessments. Residue Bridging Analysis for Seed Treatment
Uses of Neonicotinoids. 62 pp.
USEPA. 2020k. Acetamiprid: BEAD benefit assessment and response to public comments in support of registration review. 20
USEPA, PMRA, CDPR. 2014. Guidance for assessing pesticide risk to bees. Office of Chemical Safety and Pollution Prevention
Office of Pesticide Programs, Environmental Fate and Effects Division, Environmental Protection Agency, Washington
DC; Environmental Assessment Directorate, Pest Management Regulatory Agency, Health Canada, Ottawa, CN; California
Department of Pesticide Regulation.
Veres, A., Wyckhuys, K.A.G., Kiss, J., Tóth, F., Burgio, G., Pons, X., Avilla, C., Vidal, S., Razinger, J., Bazok, R., Matyjaszczyk,
E., Milosavljevi, I., Le, X.V., Zhou, W., Zhu, Z.-R., Tarno, H., Hadi, B., Lundgren, J., Bonmatin, J.-M., van Lexmond, M.B.,
Aebi, A., Rauf, A., Furlan, L. 2020. An update of the Worldwide Integrated Assessment (WIA) on systemic pesticides. Part 4:
Alternatives in major cropping systems. Environ Sci Pollut Res.
Visscher, P.K., Crailsheim, K., Sherman, G. 1996. How do honey bees (Apis mellifera) fuel their water foraging flights? Journal
of Insect Physiology 42, 1089–1094.
Whiting, S.A., Strain, K.E., Campbell, L.A., Young, B.G., Lydy, M.J. 2014. A multi-year field study to evaluate the environmental
fate and agronomic effects of insecticide mixtures. Science Total Environm. 497–498, 534–542.
Whiting, S.A. and Lydy, M.J. 2015. A site-specific ecological risk assessment for corn-associated insecticides. Integrated
Environmental Assessment and Management 11, 445-458.
Woodcock, B.A., Isaac, N.J.B., Bullock, J.M., Roy, D.B., Garthwaite, D.G., Crowe, A., Pywell, R.F. 2016. Impacts of
neonicotinoid use on long-term population changes in wild bees in England. Nature Communications 7, 12459. https://doi.
Xue, Y., Limay-Rios, V., Smith, J., Baute, T., Forero, L.G., Schaafsma, A. 2015. Quantifying Neonicotinoid Insecticide Residues
Escaping during Maize Planting with Vacuum Planters. Environ. Sci. Technol. 49, 13003–13011.
Crop names were standardized and grouped to be consistent among counties.
For the purpose of identifying areas where seed treatments might have been used, tree fruits, all vine and small fruits, and
mushrooms were excluded.
Areas in range or rangeland were excluded as we assumed no use of seed treatments in these areas.
Areas in pasture, whether irrigated or not, were assumed to be in permanent pasture, which excluded any possible seed
treatment use. This is a critical assumption as it represents a large area—almost 18 million acres according to 2016 county
reports—but probably not an unreasonable one given that we can find no registered uses of neonicotinoids on pasture
grasses. Seed treatments, however, are registered for cereals and alfalfa.
Small areas reported as “mixed cereals” (barley, oats, rye, triticale, and wheat) were assumed to be “wheat” for calculation
purposes. In fact, rates per seed of the different active ingredients are generally similar for all cereal species.
Areas recorded as being in stubble post-harvest were counted as mixed cereal.
Hay and haylage crop areas were separated into alfalfa and cereals (where seed treatments would be allowed) versus other
types (grass crops including Sudan grass, ryegrass, timothy, clover, vetch, etc.) for which we have not found any registered
seed treatment uses. Based on totals from the combined county statistics, harvested hay crops were divided almost equally
between alfalfa, cereals, and other species. Therefore, for those counties where different types of hay crops were combined,
we assigned 1/3 of the total acreage to each of these hay subtypes. For the purpose of calculations, cereal hay and haylage
were considered to be “oats.
Acreages noted to be in “volunteer” hay were excluded, as no seeding was involved.
Alfalfa seed and hay production were kept separate.
Wheat and triticale crops were combined, as they were often combined in county reports.
Sweet potato and yam crops were similarly combined.
“Vegetable vine crops” were considered to be cucurbits—specifically,cucumbers” for the purpose of calculations.
All organic crop acreages were excluded from the tallies.
Specialty Asian vegetables were usually combined in state statistics. For the purpose of calculating possible seed treatment
uses, they were considered to be leafy greens—e.g., bok choy.
For seed weight and rate calculations, the combined “herb” crop was assumed to be parsley.
For seed weight and rate calculations, crop acreages reported as “greens” were assumed to be part spinach and part kale.
“Vegetable seed production” in open fields is usually combined in statistics. Given the wide variety of possible crops and
varietals, and for the purpose of calculating seed treatment rates, squash was used as the model crop.
Asparagus are either direct seeded or started from crowns—in which case, seeds are also used, but in starting beds
elsewhere. Both planting scenarios were combined.
Roughly half of the artichoke crop is seeded annually; the other half of the crop consists of perennial plants (Smith et al.
2008). The reported artichoke areas were therefore divided in half when considering possible seed treatment use.
It was assumed that only 10% of the California rice crop was available for seed treatment use, the bulk of the crop being
water seeded (paddy rice). Neonic seed treatments are not registered for paddy seeding.
Crop or crop
grouping Acetamiprid Clothianidin Imidacloprid Thiamethoxam All neonics
alfalfa 1 1
amaranth 3 1 4
arrowroot 2 2
artichoke 1 1
Asian vegetables 1 2 16 19
barley 8 45 7 60
bean 4 42 122 168
borage 13 2 15
broccoli 3 3 6
Brussel sprouts 1 1
buckwheat 6 1 7
cabbage 1 13 1 15
canola 2 6 43 6 57
cardoon 1 1
carrot 2 17 1 20
cassava 1 1
cauliflower 1 1
celery 1 1
chard 1 1
chicory 3 1 4
corn 15 41 9 65
cotton 1 25 4 30
crambe 2 2
cucumber 3 3
endive 3 1 4
fennel 1 1
flax 13 2 15
Edible flowers 10 1 11
ginger 1 1
greens 18 1 11 30
herbs 6 2 8
22 is is a more granular grouping of crops than the legal framework provided by USEPA and recognized by California (Appendix 4.40 CFR 180.41).
Crop or crop
grouping Acetamiprid Clothianidin Imidacloprid Thiamethoxam All neonics
kale 1 1
kohlrabi 1 1
leek 3 3
lentil 2 6 8
lettuce 13 4 17
lupin 8 8 16
melon 14 14
millet 8 10 2 20
mustard 1 18 4 23
oats 8 35 5 48
onion 6 1 7
pea 30 55 85
peanuts 1 3 4
potato 1 3 10 5 19
pumpkin 1 1
rhubarb 1 1
rice 2 2
rye 6 24 5 35
safflower 11 2 13
sorghum 2 25 2 29
sorrel 3 1 4
soybean 2 28 15 45
spinach 12 3 15
squash 13 13
sugar beet 4 24 1 29
sunflower 5 2 7
sweet corn 3 24 1 28
sweet potato 2 2
taro 1 1
turmeric 1 1
wheat 14 82 16 112
wild rice 1 1
For certain crop groups with many varietals separately mentioned on labels, a high number of tabulated listings may be partly
artefactual. For example, the “bean” group includes adzuki bean, wax bean, kidney bean, etc., and the table shows how often
“bean” is listed on a label. Similarly, issues may arise with the number of individual specialty crop species amalgamated into
larger groups (e.g., greens, herbs, or Asian vegetables). Nevertheless, the tabulation shows the sheer number and diversity of
approved seed treatment uses and the incredible penetration of neonics into the agricultural seed treatment business.
region County
Potential seed treatment use PUR data Proportional
increase from
seed treatmentsb
lbs a.i.
lbs a.i.
lbs a.i.
lbs a.i.
lbs a.i.
lbs a.i.
Central coast
Alameda 12 103 58 0 461 13 1.37
Costa 846 789 773 69 1,284 48 2.72
Lake 1 20 11 0 420 0 1.08
Marin 88 86 98 0 0 0 273
Monterey 54,146 1,606 53,185 6,538 18,975 10,287 4.04
Napa 3 7 7 242 1,309 812 1.01
San Benito 1,132 647 1,847 278 1,633 300 2.64
San Luis
Obispo 1,752 1,349 2,142 143 12,962 642 1.38
San Mateo 68 76 73 0 73 55 2.70
Santa Clara 205 188 276 416 572 340 1.50
Santa Cruz 1,113 22 956 74 493 137 3.97
Sonoma 129 1,288 584 239 2,723 172 1.64
Lassen 138 2,624 1,481 0 0 0 4,240
Modoc 755 3,177 3,157 0 139 6 50.0
Plumas 5 48 55 0 0 0 109
Del Norte 3 56 30 0 0 0 90.1
Humboldt 23 177 111 0 0 0 312
Mendocino 14 124 80 0 516 8 1.42
Butte 102 810 1,935 89 1,060 1 3.48
Colusa 418 3,892 4,097 5 1,170 2 8.14
Glenn 1,347 2,625 3,087 2 1,467 0 5.81
Sacramento 2,568 4,859 3,401 171 10,255 407 2.00
Solano 805 5,907 2,685 163 1,248 30 7.52
Sutter 1,093 2,923 3,800 60 749 0 10.7
Tehama 159 1,458 615 7 101 0 21.8
Yolo 888 7,724 4,245 32 3,116 150 4.90
Yuba 5 63 687 25 278 0 3.49
region County
Potential seed treatment use PUR data Proportional
increase from
seed treatmentsb
lbs a.i.
lbs a.i.
lbs a.i.
lbs a.i.
lbs a.i.
lbs a.i.
San Joaquin
Fresno 8,749 8,886 12,759 4,193 69,604 6,218 1.38
Kern 10,229 7,950 8,318 2,140 50,131 2,855 1.48
Kings 4,291 9,815 8,555 1,185 14,891 1,494 2.29
Madera 1,449 1,506 1,289 549 17,359 512 1.23
Merced 14,584 11,974 9,687 1,467 11,861 867 3.55
San Joaquin 7,589 14,475 9,517 251 18,316 1,788 2.55
Stanislaus 1,850 3,833 2,125 190 5,769 276 2.25
Tulare 13,319 8,579 10,205 1,597 30,358 7,846 1.81
Amador 4 59 27 0 33 0 3.77
Calaveras 5 14 22 0 9 0 5.58
El Dorado 5 5 8 18 67 0 1.20
Inyo 2 34 18 0 0 0 53.9
Mariposa 1 12 6 0 0 0 19.4
Mono 6 119 63 0 0 0 189
Nevada 9 12 20 0 1 0 63.0
Placer 14 152 261 0 69 0 7.20
Sierra 1 20 17 0 0 0 40.0
Tuolomne 0 0 0 0 0 0 1.00
Shasta 56 627 1,715 4 6 7 137
Siskiyou 743 1,147 1,796 0 202 18 17.8
Trinity 1 5 5 0 0 0 11.9
Imperial 22,605 7,073 22,117 626 21,225 664 3.30
Los Angeles 13 85 64 0 735 20 1.21
Orange 23 0 16 0 213 55 1.15
Riverside 1,800 4,263 4,187 17 9,586 6 2.07
Bernardino 129 530 746 0 643 77 2.95
San Diego 88 115 216 0 2,717 182 1.14
Barbara 7,895 401 6,717 1,022 24,720 1,266 1.56
Ventura 1,768 348 13,528 313 6,948 4,176 2.37
Unspecified 463 13,400 4,827
a CLO= clothianidin; IMI= imidacloprid; THI=thiamethoxam.
b To avoid division by 0, reported PUR uses changed from 0 to 1 lb.
Corn is known to extensively use neonic seed treatments. Douglas and Tooker (2015) estimated that, by 2011, between 79%
and 100% of the corn acreage nationwide used neonic seed treatments and found evidence that the rate was still increasing. By
now, most of the corn crop area is expected to be treated with either clothianidin or thiamethoxam. Indeed, it may be difficult
for growers to find untreated seed to plant should they wish to do so. Lack of crop rotation remains the main impediment for
moving away from prophylactic control of corn rootworm (Veres et al. 2020).
The bulk of neonic use in cotton comes from seed treatments. The number of seed treatment applications slightly exceeded
the number of base acres—1.04 to 1—when assessed (USEPA 2017). This suggests that all cotton acres receive a seed
treatment, primarily thiamethoxam and imidacloprid.
Wheat and other cereal crops
USEPA (2018) estimated that 20% of winter wheat and 27% of spring wheat received a neonic seed treatment. However, these
data date from the period 2010-2014 and likely are underestimates of current conditions.
Because aphids in small grain crops including cereals can vector diseases, extension recommendations often advise using
neonic seed treatments for early season control of aphids (USEPA 2018). In comments sent to USEPA during the current
review of neonicotinoid benefits, a representative of the National Barley growers association argued against possible
restrictions of neonic use (USEPA 2019b). Similarly, the National Association of Wheat growers argued that seed treatments
provided the most efficient treatment of early season pests (USEPA 2019b).
Sugar beets
As is the case with small grain crops, the possibility of vector-borne disease in sugar beets means that the use of seed
treatments is often recommended (USEPA 2019b). Based on data collected between 2010 and 2014, USEPA (2018) estimated
that 46% of the crop acreage nationally received a seed treatment.
Small grain and oilseed crops
USEPA (2019b) estimated that 44% of the sorghum crop received a neonic seed treatment. Once again, this was based on
now-outdated information (2010-2014).
All three active ingredients are registered for oilseeds; a seed treatment is now considered the “standard” in canola as judged by
the many articles looking into its use. No information could be obtained on sorghum, flax, safflower, and sunflower.
Vegetable crops
USEPA (2018) reviewed seed treatment uses of neonics in vegetable crops. Of most concern when considering seed treatments
are crops that are direct seeded given that transplants are more likely to be grown under controlled conditions—e.g., in a
greenhouse. Depending on how greenhouse effluents are collected and treated, greenhouse use may also pose contamination
concerns, particularly for water contamination—but likely from a point source rather than diffuse sources.
In its benefits review of neonics, USEPA (2018) had this to say about vegetable crops:
BEAD does not have data on the use of neonicotinoid seed treatments in vegetable crops and thus sought stakeholder feedback
concerning usage. Per the American Seed Trade Association [ASTA], less than 15% of vegetable acreage is estimated to be
planted with neonicotinoid seed treatments based on best professional judgement (ASTA pers. comm. 2018); ASTA did not
provide estimates by individual crops. ASTA reported that clothianidin and thiamethoxam are preferred by growers over
imidacloprid seed treatments.
USEPA (2018) believed that there was a greater reliance on soil-applied insecticides (granules or drenches) than seed
treatments because the former (presumably drenches) may better target above ground pests. However, they also acknowledged
that several extension services recommend neonic-treated seeds. Further, it is likely that upcoming possible restrictions on the
foliar use of neonics (e.g., in cucurbits) will result in a shift to more seed treatment uses.
I attempted to verify the 15% estimate by means of the differential 2014/2015 use reported from commercial surveys and
made public by USGS. USGS reports the use of the main nitroguanidine neonics separately on “vegetable and fruit” crops.
Excluding California (because seed treatment use was not estimated in either year), the total quantity of use dropped by 14%
between 2014 and 2015. Assuming negligible use of seed treatments in fruit production, this drop may be ascribed to vegetable
production and suggests that the ASTA 15% total use estimate appears to be reasonable. However, reliance on the USGS-
reported differential for 2014/15 assumes that the Kynetec commercial surveys had a good handle on seed treatment uses in
2014. They stopped collecting the information in 2015, in part, because of the increased difficulty of acquiring reliable data
from users (Hitaj et al. 2020).
It is clear also that seed treatment use will vary widely among the many vegetable crops. Whitefly and thrips are key targets of
neonic seed treatments in vegetable production. Leafy vegetables and green onions are especially likely to be treated given that
an economic threshold has not been established (USEPA 2018). Additionally, there is wide resistance to most other insecticide
groups in those crops.
A study proposing to look at the fate of neonics applied as seed treatments to lettuce (USGS 2020) stated unambiguously that
in California, lettuce is grown from neonicotinoid treated seeds.” There appears to be no doubt in the mind of these researchers
that use of seed treatments in that crop is the norm rather than a 15% possibility.
A recent development is the appearance of a new pest, the introduced bagrada bug (Bagrada hilaris - a species of stink bug).
It is considered to be problematic for cole crops especially. The pest has now expanded from southern California to the entire
state making the use of neonic seed treatments much more likely in broccoli, cauliflower, and cabbage. When its favored plants
are not present, it attacks a wide range of other crop species (bell pepper, melon, papaya, tomato, and capers, corn, Sudan
grass, sorghum, sunflowers, potato, cotton, and some legumes, including snap beans) (Reed et al. 2014).
Finally, USEPA (2018) proposes that neonic seed treatments may be more important in low-acreage vegetable crops for
which other options have not been well developed. Keeping all of this in mind, it is clear that 15% should be considered a low
estimate and that usage is probably increasing steadily in many vegetable crops.
It is unclear where potato “seeds” are treated—and if treated on farm, whether this is recorded in PUR data. In response to
EPA request for comments, the National Potato Council and other state organizations stressed the importance of neonics as
seed treatments in potato culture (USEPA 2019b). The first uses of neonicotinoids were in potato.
Most of the rice in California is water-seeded. Seed treatments are only registered for use in dry-seeded systems and are
therefore of limited or no utility in California. Elsewhere, an estimated 45-75% of the rice is treated with a seed treatment
(USEPA 2019c). This prohibition, interestingly, does not seem to apply to wild rice. Our understanding is that the crop is also
seeded by air into paddies. This grass crop grows in wetlands. However, thiamethoxam is registered for use. It is unclear why it
is registered if this crop is indeed wet seeded.
A key uncertainty in our assessment is the issue of neonic seed treatments for forages. Alfalfa and cereal forages, specifically,
make up a large proportion (23.7%) of the total crop area on which crops could be planted with the use of seed treatments.
This excludes 1/3 of the forage area planted to grasses and other species for which seed treatments are not registered, such as
“pasture and hay, for which USGS 2014 estimates reported no use of neonics.
For alfalfa (presumably grown for seed), the amounts reported by USGS (from PUR data) were identical in 2014 and 2015
suggesting no or minimal seed treatment use. Only thiamethoxam was reported as being used at all. The use of thiamethoxam
on treated seed is likely to increase given that they were only introduced in time for 2013 plantings (AG Professional
2013), although neonic seed treatments were already registered for cereal use. According to AG professional (2013), the
thiamethoxam component of the formulated seed treatment is able to help alfalfa crops develop stronger roots, use inputs more
efficiently; emerge faster, and grow more evenly, even in the absence of insects (emphasis added).
It isn’t known at this stage whether we will see the same exponential growth in alfalfa that was seen in corn, soy, and cereals. It
also is not clear whether any of the cereal hay and haylage is grown with the use of the many available seed treatments.
district County
Sum of planted
Sum of possible
treated acreage
Proportion with
possible seed treatment
Central coast
Total - all counties 511,155 462,742 91%
Alameda 4,377 3,342 76%
Contra costa 26,623 21,965 83%
Lake 1,282 856 67%
Marin 4,769 3,359 70%
Monterey 344,009 337,438 98%
Napa 475 311 65%
San Benito 36,547 20,894 57%
San Luis Obispo 39,405 38,138 97%
San Mateo 2,735 2,218 81%
Santa Clara 16,497 7,479 45%
Santa Cruz 5,675 5,095 90%
Sonoma 28,762 21,648 75%
Total - all counties 165,643 130,587 79%
Lassen 73,442 63,667 87%
Modoc 85,566 61,288 72%
Plumas 6,634 5,631 85%
Northern coast
Total - all counties 22,166 14,821 67%
Del Norte 3,513 2,343 67%
Humboldt 10,859 7,272 67%
Mendocino 7,794 5,206 67%
Total - all counties 1,267,490 660,722 52%
Butte 107,061 20,708 19%
Colusa 223,016 76,999 35%
Glenn 159,144 75,481 47%
Sacramento 129,877 113,442 87%
Solano 126,267 111,434 88%
Sutter 189,108 68,073 36%
Tehama 27,859 23,154 83%
Yolo 262,330 164,350 63%
Yuba 42,828 7,082 17%
district County
Sum of planted
Sum of possible
treated acreage
Proportion with
possible seed treatment
San Joaquin
Total - all counties 2,465,333 1,968,959 80%
Fresno 437,493 289,337 66%
Kern 253,893 232,180 91%
Kings 332,773 257,386 77%
Madera 76,116 49,463 65%
Merced 370,133 306,500 83%
San Joaquin 341,058 310,159 91%
Stanislaus 232,374 159,316 69%
Tulare 421,494 364,620 87%
Sierra mountains
Total - all counties 35,071 16,628 47%
Amador 3,108 1,869 60%
Calaveras 888 584 66%
El dorado 326 215 66%
Inyo 2,087 1,391 67%
Mariposa 726 484 67%
Mono 7,378 4,919 67%
Nevada 796 536 67%
Placer 17,598 4,892 28%
Sierra 2,164 1,739 80%
Total - all counties 104,697 85,684 82%
Shasta 36,073 22,981 64%
Siskiyou 68,295 62,486 91%
Trinity 329 217 66%
Total - all counties 531,425 509,531 96%
Imperial 236,768 236,768 100%
Los Angeles 5,748 4,023 70%
Orange 201 184 92%
Riverside 132,102 126,306 96%
San Bernardino 21,642 20,167 93%
San Diego 5,825 3,318 57%
Santa Barbara 86,712 81,814 94%
Ventura 42,427 36,951 87%
Unspecified Unspecified 150,352 148,785 99%
California Total Grand total 5,253,331 3,998,459 76%
Figure 20 from EPA 2020j (Attachment 4) summarizing industry data on mean pollen residues in three crops following the
use of seed treatments. Residue levels are normalized to 0.1 mg a.i./seed. Open circles indicate that none of the samples had
levels above detection. Crosses indicate that at least one sample was below the limit of detection.
Figure 21 from EPA 2020j (Attachment 4) summarizing industry data on mean nectar residues in four crops following the use
of seed treatments. Residue levels are normalized to 0.1 mg a.i./seed. Open circles indicate that none of the samples had levels
above detection. Crosses indicate that at least one sample was below the limit of detection.
Full-text available
Pesticides usually occur as mixtures of multiple chemicals in the natural aquatic ecosystem, so research based on the toxicity data of a single compound on aquatic organisms is not enough to accurately assess the actual toxicity risk of pesticides. There is still a gap in the research on the reproductive toxicity of combined insecticides, herbicides and fungicides on zebrafish (Danio rerio). In this study, zebrafish were used to systematically investigate the separate and combined reproductive toxicity of imidacloprid (IMI), acetochlor (ACT) and tebuconazole (TBZ), which are commonly used in rice fields. Adult zebrafish were exposed to the three pesticides individually and in combination for 28 days, and the number, heartbeat, deformation rate, body length, and swim bladder development of F1 offspring embryos were observed and the reproductive hormones testosterone (T), estradiol (E2), and vitellogenin (VTG) contents and the expressions of nine reproductive genes (ar, esr2a, vtg1, gr, star, fshr, hmgcrb, 3βhsd and vasa) in the testes of the male and the ovaries of the female F0 zebrafish adults were measured to evaluate the individual and combined effects. The results showed that exposure to the mixtures of IMI, ACT and TBZ resulted in a decrease in heartbeat, body length and swim bladder development and an increase in the deformity rate of F1 offspring embryos compared to the individual exposure groups. In the combined exposure group, the content of T decreased significantly and the content of VTG increased significantly in the testes of the males; the content of T significantly increased, while the content of E2 and VTG significantly decreased in the ovaries of the females, indicating that combined exposure showed a more obvious endocrine-disrupting effect compared to the individual exposures. In addition, the expression of nine reproductive genes was significantly altered compared to the individual exposure groups. Therefore, our results indicated that the mixture of IMI, ACT and TBZ caused fewer number of F1 embryos, higher developmental defects of F1, greater disruption in the content of reproductive hormones and the expression of reproductive genes compared to the individual pesticides at the corresponding doses. Therefore, the presence of pesticides in mixtures in the real water environment is likely to increase the toxic reproductive effects on zebrafish and cause more serious impacts on aquatic ecosystems.
Full-text available
In March 1972, Frederick Coulston and colleagues at the Albany Medical College reported results of an intentional chlorpyrifos dosing study to the study’s sponsor, Dow Chemical Company. Their report concluded that 0.03 mg/kg-day was the chronic no-observed-adverse-effect-level (NOAEL) for chlorpyrifos in humans. We demonstrate here that a proper analysis by the original statistical method should have found a lower NOAEL (0.014 mg/kg-day), and that use of statistical methods first available in 1982 would have shown that even the lowest dose in the study had a significant treatment effect. The original analysis, conducted by Dow-employed statisticians, did not undergo formal peer review; nevertheless, EPA cited the Coulston study as credible research and kept its reported NOAEL as a point of departure for risk assessments throughout much of the 1980′s and 1990′s. During that period, EPA allowed chlorpyrifos to be registered for multiple residential uses that were later cancelled to reduce potential health impacts to children and infants. Had appropriate analyses been employed in the evaluation of this study, it is likely that many of those registered uses of chlorpyrifos would not have been authorized by EPA. This work demonstrates that reliance by pesticide regulators on research results that have not been properly peer-reviewed may needlessly endanger the public.
Full-text available
We present a synthetic review and expert consultation that assesses the actual risks posed by arthropod pests in four major crops, identifies targets for integrated pest management (IPM) in terms of cultivated land needing pest control and gauges the implementation “readiness” of non-chemical alternatives. Our assessment focuses on the world’s primary target pests for neonicotinoid-based management: western corn rootworm (WCR, Diabrotica virgifera virgifera) in maize; wireworms (Agriotes spp.) in maize and winter wheat; bird cherry-oat aphid (Rhopalosiphum padi) in winter wheat; brown planthopper (BPH, Nilaparvata lugens) in rice; cotton aphid (Aphis gossypii) and silver-leaf whitefly (SLW, Bemisia tabaci) in cotton. First, we queried scientific literature databases and consulted experts from different countries in Europe, North America, and Asia about available IPM tools for each crop-pest system. Next, using an online survey, we quantitatively assessed the economic relevance of target pests by compiling country-level records of crop damage, yield impacts, extent of insecticide usage, and “readiness” status of various pest management alternatives (i.e., research, plot-scale validation, grower-uptake). Biological control received considerable scientific attention, while agronomic strategies (e.g., crop rotation), insurance schemes, decision support systems (DSS), and innovative pesticide application modes were listed as key alternatives. Our study identifies opportunities to advance applied research, IPM technology validation, and grower education to halt or drastically reduce our over-reliance on systemic insecticides globally.
Full-text available
Our mechanistic understanding of the toxicity of chemicals that target biochemical and/or physiological pathways, such as pesticides and medical drugs is that they do so by binding to specific molecules. The nature of the latter molecules (e.g., enzymes, receptors, DNA, proteins, etc.) and the strength of the binding to such chemicals elicit a toxic effect in organisms, which magnitude depends on the doses exposed in a given timeframe. While dose and time of exposure are critical factors determining the toxicity of pesticides, different types of chemicals behave differently. Experimental evidence demonstrates that the toxicity of neonicotinoids increases with exposure time as much as with the dose, and therefore it has been described as time-cumulative toxicity. Examples for aquatic and terrestrial organisms are shown here. This pattern of toxicity, also found among carcinogenic compounds and other toxicants, has been ignored in ecotoxicology and risk assessments for a long time. The implications of the time-cumulative toxicity of neonicotinoids on non-target organisms of aquatic and terrestrial environments are far reaching. Firstly, neonicotinoids are incompatible with integrated pest management (IPM) approaches and secondly regulatory assessments for this class of compounds cannot be based solely on exposure doses but need also to take into consideration the time factor.
Full-text available
Agricultural soil pests, including wireworms (Coleoptera: Elateridae), are managed primarily with pesticides applied directly to seeds before sowing. Seeds coated with neonicotinoids have been used widely in Quebec (Canada) for several years. To assess the agronomic and economic value of neonicotinoid seed treatments in soybeans and corn in Quebec, trials were conducted from 2012 to 2016 in 84 fields across seven regions in Quebec. We evaluated the effect of neonicotinoid seed treatments on soil pest densities, crop damage and yield. The results showed that 92.6% of corn fields and 69.0% of soybean fields had less than 1 wireworm per bait trap. However, no significant differences in plant stand or yield were observed between treated and untreated corn or soybeans during the study. This study shows that neonicotinoid seed treatments in field crops in Quebec are useful in less than 5% of cases, given the very low level of pest-associated pressure and damage, and that they should not be used prophylactically. Integrated pest management (IPM) strategies need to be developed for soil insect pests to offer effective alternative solutions to producers.
Technical Report
Full-text available
A risk assessment of bats and neonicotinoid insecticides.
Full-text available
Neonicotinoid seed treatments are extensively used to systemically protect corn from invertebrate herbivory. Interseeding cover crops can promote beneficial insect communities and their ecosystem services such as predation on pests, and this practice is gaining interest from farmers. In this study, cereal rye (Secale cereale) and hairy vetch (Vicia villosa) were planted between rows of early vegetative corn that had been seed-treated with thiamethoxam. Thiamethoxam and its insecticidal metabolite, clothianidin were quantified in cover crop leaves throughout the growing season. Thiamethoxam was present in cereal rye at concentrations ranging from 0 to 0.33 ± 0.09 ng/g of leaf tissue and was detected on six out of seven collection dates. Cereal rye leaves contained clothianidin at concentrations from 1.05 ± 0.22 to 2.61 ± 0.24 ng/g and was present on all sampling dates. Both thiamethoxam and clothianidin were detected in hairy vetch on all sampling dates at rates ranging from 0.10 ± 0.05 to 0.51 ± 0.11 ng/g and 0.56 ± 0.15 to 9.73 ± 5.04 ng/g of leaf tissue, respectively. Clothianidin was measured at a higher concentration than its precursor, thiamethoxam, in both plant species on every sampling date. Neonicotinoids entering interseeded cover crops from adjacent treated plants is a newly discovered route of exposure and potential hazard for non-target beneficial invertebrates. Future research efforts should examine the effects of systemic insecticides on biological communities in agroecosystems whose goal is to diversify plant communities using methods such as cover cropping.
Full-text available
Authorization of pesticides for market release requires toxicity testing on animals, typically performed by test laboratories on contract with the pesticide producer. The latter provides the results and summary to the regulatory authorities. For the commonly used pesticide chlorpyrifos, an industry-funded toxicity study concludes that no selective effects on neurodevelopment occur even at high exposures. In contrast, the evidence from independent studies points to adverse effects of current exposures on cognitive development in children. We reviewed the industry-funded developmental neurotoxicity test data on chlorpyrifos and the related substance chlorpyrifos-methyl. We noted treatment-related changes in a brain dimension measure for chlorpyrifos at all dose levels tested, although not been reported in the original test summary. We further found issues which inappropriately decrease the ability of the studies to reveal true effects, including a dosage regimen that resulted in too low exposure of the nursing pups for chlorpyrifos and possibly for chlorpyrifos-methyl, and a failure to detect any neurobehavioral effects of lead nitrate used as positive control in the chlorpyrifos study. Our observations thus suggest that conclusions in test reports submitted by the producer may be misleading. This discrepancy affects the ability of regulatory authorities to perform a valid and safe evaluation of these pesticides. The difference between raw data and conclusions in the test reports indicates a potential existence of bias that would require regulatory attention and possible resolution. Electronic supplementary material The online version of this article (10.1186/s12940-018-0421-y) contains supplementary material, which is available to authorized users.
Farmers, regulators, and researchers rely on pesticide use data to assess the effects of pesticides on crop yield, farm economics, off-target organisms, and human health. The publicly available pesticide use data in the United States do not currently account for pesticides applied as seed treatments. We find that seed treatment use has increased in major field crops over the last several decades but that there is a high degree of uncertainty about the extent of acreage planted with treated seeds, the amount of regional variability, and the use of certain active ingredients. One reason for this uncertainty is that farmers are less likely to know what pesticides are on their seed than they are about what pesticides are applied conventionally to their crops. This lack of information affects the quality and availability of seed treatment data and also farmers’ ability to tailor pesticide use to production and environmental goals.
Native bees are in decline as many species are sensitive to habitat loss, climate change, and non-target exposure to synthetic pesticides. Recent laboratory and semi-field assessments of pesticide impacts on bees have focused on neonicotinoid insecticides. However, field studies evaluating influences of neonicotinoid seed treatments on native bee communities of North America are absent from the literature. On four Conservation Areas of Missouri, we sampled row-cropped (treated, n = 15) and reference (untreated, n = 9) agricultural fields, and their surrounding field margins for neonicotinoids in soil and non-target vegetation (i.e., native wildflowers). Wildflowers were further collected and screened for the presence of fungicides. Concurrently, we sampled native bees over three discrete time points throughout the agricultural growing season to assess potential impacts of seed treatment use on local bee populations over time. Neonicotinoids were detected in 87% to 100% of treated field soils and 22% to 56% of reference field soils. In adjacent field margin soils, quantifiable concentrations were measured near treated (53% to 93% detection) and untreated fields (33% to 56% detection). Fungicides were detected in < 40% of wildflowers, whereas neonicotinoids were rarely detected in field margin vegetation (< 7%). Neonicotinoid concentrations in margin soils were negatively associated with native bee richness (β = −0.21, P < 0.05). Field margins with a combination of greater neonicotinoid concentrations in soil and fungicides in wildflowers also contained fewer wild bee species (β = −0.21, P < 0.001). By comparison, bee abundance was positively influenced by the number of wildflower species in bloom with no apparent impact of pesticides. Results of this study indicate that neonicotinoids in soil are a potential route of exposure for pollinator communities, specifically ground-nesting species. Importantly, native bee richness in non-target field margins may be negatively affected by the use of neonicotinoid seed treatments in agroecosystems.
Neonicotinoid insecticides are frequently detected in surface waters near agricultural areas, leading to a potential for chronic exposure to sensitive aquatic species. The midge Chironomus dilutus and the mayfly Neocloeon triangulifer have been shown to be acutely sensitive to neonicotinoids. While previous studies have established chronic effects of some neonicotinoids on C. dilutus, reproduction remains unstudied. Toxic effects have not been assessed using N. triangulifer. Here we present results of chronic, static‐renewal tests for 6 neonicotinoids (acetamiprid, clothianidin, dinotefuran, imidacloprid, thiacloprid, and thiamethoxam) with C. dilutus (≤56‐d in length) and N. triangulifer (≤32‐d in length). Emergence was generally the most sensitive endpoint for both species across all neonicotinoids. EC10s (emergence) were 0.03–1.1 µg L−1 for acetamiprid, clothianidin, imidacloprid, and thiacloprid. Dinotefuran and thiamethoxam were less potent, with EC10s (C. dilutus) or EC50s (N. triangulifer) of 2.2–11.2 µg L−1. Hazard was assessed through comparison of neonicotinoid environmental concentrations from agricultural surface waters in Ontario (Canada) with either the 5th percentile hazard concentration (for imidacloprid) or species specific EC10s from the present study (for all remaining neonicotinoids). Resulting hazard quotients indicated little to no hazard (HQ < 1) in terms of chronic toxicity for acetamiprid, dinotefuran, thiacloprid, or thiamethoxam. A moderate hazard (HQ > 1) was found for emergence of N. triangulifer for clothianidin, and a high hazard (HQ = 74) was found for imidacloprid. This article is protected by copyright. All rights reserved