Impact of Bt corn pollen on monarch butterfly
populations: A risk assessment
Mark K. Sears*†, Richard L. Hellmich‡, Diane E. Stanley-Horn*, Karen S. Oberhauser§, John M. Pleasants¶,
Heather R. Mattila*, Blair D. Siegfried?, and Galen P. Dively**
*Department of Environmental Biology, University of Guelph, Guelph, ON, Canada N1G 2W1;‡United States Department of Agriculture–Agricultural
Research Service, Corn Insects and Crop Genetics Research Unit and Department of Entomology, Iowa State University, Ames, IA 50011;
§Department of Ecology and Evolutionary Biology, University of Minnesota, St. Paul, MN 55108;¶Department of Zoology and
Genetics, Iowa State University, Ames, IA 50011;?Department of Entomology, University of Nebraska, Lincoln, NE 68583;
and **Department of Entomology, University of Maryland, College Park, MD 20742
Edited by M. R. Berenbaum, University of Illinois at Urbana–Champaign, Urbana, IL, and approved August 17, 2001 (received for review June 28, 2001)
A collaborative research effort by scientists in several states and in
Canada has produced information to develop a formal risk assess-
ment of the impact of Bt corn on monarch butterfly (Danaus
plexippus) populations. Information was sought on the acute toxic
effects of Bt corn pollen and the degree to which monarch larvae
would be exposed to toxic amounts of Bt pollen on its host plant,
the common milkweed, Asclepias syriaca, found in and around
cornfields. Expression of Cry proteins, the active toxicant found in
Bt corn tissues, differed among hybrids, and especially so in the
concentrations found in pollen of different events. In most com-
mercial hybrids, Bt expression in pollen is low, and laboratory and
field studies show no acute toxic effects at any pollen density that
would be encountered in the field. Other factors mitigating expo-
sure of larvae include the variable and limited overlap between
pollen shed and larval activity periods, the fact that only a portion
of the monarch population utilizes milkweed stands in and near
American corn-growing areas. This 2-year study suggests that the
impact of Bt corn pollen from current commercial hybrids on
monarch butterfly populations is negligible.
(Bt) arose after the publication by Losey et al.(1) on the potential
risk of corn pollen expressing lepidopteran-active Cry protein to
the monarch butterfly, Danaus plexippus L. However, the U.S.
Environmental Protection Agency (EPA) concluded in an ear-
lier report that the potential impact of Bt corn pollen, which
contains variable amounts of Cry protein, on sensitive larvae of
Lepidoptera was negligible because of factors that limit envi-
pollen to monarch butterflies can now be undertaken because of
the data reported in this issue of PNAS (3–6) that address
exposure and toxic effects of Bt corn pollen.
The research contributions reported here represent a collab-
orative effort established to specifically address the question of
risk associated with Bt corn pollen to the monarch butterfly. In
December 1999, the EPA issued a data call-in requesting indus-
try, researchers and all interested parties to submit information
and comments by March 2001 for use in evaluation and potential
reregistration of corn hybrids containing Cry proteins (http:??
The U.S. Department of Agriculture Agricultural Research
Service (USDA–ARS) sponsored a Monarch Research Work-
shop in February 2000 to identify research priorities regarding Bt
corn and monarch butterflies, to establish cooperation among
researchers, and to respond to the EPA request for data. A
request for proposals based on workshop priorities was an-
nounced in April, after which a steering committee, including
Adrianna Hewings (USDA–ARS), Eldon Ortman (Purdue Uni-
versity), Mark Scriber (Michigan State University), Eric Sachs
(Monsanto), and Margaret Mellon (Union of Concerned Scien-
tists) selected projects to be funded. Funding came from a grant
oncern regarding nontarget effects of transgenic crops con-
taining transgenes from the organism Bacillus thuringiensis
pool provided by ARS and the Agricultural Biotechnology
Stewardship Technical Committee (7). The guiding principles
for problem formulation followed by the consortium were the
elements of risk assessment that underlie the approach by EPA
to ecological risk assessment (http:??www.epa.gov?NCEA?
Only three papers concerning the impact of Bt corn pollen on
nontarget Lepidoptera have been published (1, 8, 9), and they
are limited in their application to risk assessment (7). For
example, the dose of pollen was not specified in the exposure
study by Losey et al. (1), and the study by Jesse and Obrycki (8)
used pollen collection and handling techniques that probably
resulted in contamination from corn anthers or tassel fragments,
which contain significantly higher levels of Cry protein than the
pollen (3). Finally, neither study addressed the spatial or tem-
poral potential for exposure by monarch larvae to pollen in
cornfields, thereby precluding a risk assessment.
risk of exposure of monarch larvae to Bt corn pollen and the
in eastern North America by using recently published informa-
tion based on collaborative research by scientists in the U.S. and
Canada (3–6). We use an approach to risk assessment that has
been performed for many nontarget species in relation to
pesticides (10–14), industrial by-products (15, 16), and other
potential toxicants found in the environment (17). The approach
to this process is consistent, well documented, and standardized
(http:??www.epa.gov?NCEA?ecorsk.htm). It requires consider-
ation of both the expression of toxicity and the likelihood of
exposure to the toxicant as the basic components for a risk
Materials and Methods
Hazard Identification. Toxicity of purified Bt proteins to larval
stages of butterflies and moths is well known (18, 19). Studies
conducted on the use of Bt sprays in forests for gypsy moth
control have shown that Cry proteins can adversely affect
nontarget Lepidoptera (20, 21). Field data from these studies
indicated a temporary reduction in lepidopteran populations
during prolonged Bt use, although widespread irreversible harm
was not apparent (22). Lepidopteran-active Bt protein expressed
in pollen of Bt corn hybrids may pose a risk to sensitive species,
such as monarch butterflies, in or near cornfields during anthesis
(1, 8). Milkweeds, Asclepias spp., and especially common milk-
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: Bt, Bacillus thuringiensis; USDA–ARS, U.S. Department of Agriculture Ag-
ricultural Research Service; EPA, U.S. Environmental Protection Agency; LOEC, lowest-
†To whom reprint requests should be addressed. E-mail: email@example.com.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
October 9, 2001 ?
vol. 98 ?
no. 21 ?
weed, Asclepias syriaca (L.), are the sole larval food source for
monarch butterfly larvae and are abundant throughout the
corn-growing regions of North America (23). As such, hazard
from Bt corn pollen deposited on milkweed leaves warrants
consideration of its ecological risk to monarch populations.
Conceptual Model. Risk assessment requires knowledge of four
essential components: (i) hazard identification, (ii) nature of dose
and (iv) characterization of risk (24). Components of a risk assess-
ment approach as applied to the case of Bt corn and monarch
butterfly are depicted in Fig. 1. Bt proteins expressed in corn plant
tissues can bring about specific reactions in the gut of lepidopteran
The magnitude of the reaction will depend on the type of protein
of protein expressed in pollen grains from different events, the
amount of pollen consumed by larvae of different developmental
stages, and the susceptibility of larvae to the Bt protein. That a
of toxic effects is necessary to establish the first component of risk.
thresholds for comparison against the dose encountered within the
environment. These toxicity thresholds will vary based on expres-
sion levels for individual Bt corn events in conjunction with
environmental factors determining ecological exposure.
Consideration of risk as a function of exposure and effect
requires that lines of evidence be established in four areas of
inquiry: (i) Is there some density of Bt pollen on milkweed leaves
that represents a lethal or sublethal threat to monarch larvae or
later stages of development? (ii) What proportion of Bt pollen
deposited on milkweed leaves in and around cornfields exceeds the
toxicity threshold for larvae of monarchs? (iii) What proportion of
monarch populations use milkweed in and near cornfields? (iv)
What is the degree of overlap between the phenological stages of
monarch larvae and corn anthesis over the shared range of these
summarized in the four companion papers (3–6).
Characterization of Effects of Bt Corn Pollen. The Cry1A proteins
expressed in most commercial Bt corn hybrids are toxic to the
monarch butterfly (3). Mortality, expressed as LD50, was esti-
mated at 3.3 ng of protein?ml diet, whereas growth inhibition
(EC50) was estimated to be 0.76 ng?ml (2). However, the
expression of Cry1Ab endotoxin within pollen of various events
varies considerably depending on the promoter gene involved
(26). Expression is greatest in event 176 Bt corn (1.1–7.1 ?g?gm
event exceeds, by nearly two orders of magnitude, protein
expression in events Bt11 and Mon810 (0.09 ?g?gm pollen)
(http:??www.epa.gov?scipoly?sap?2000?october?), which is
near the current level of detection by immunoassay.
Laboratory bioassays of pollen fed to first instar monarchs for
4 days on leaf discs or whole detached leaves of common
milkweed, A. syriaca, indicate that pollen from event 176 Bt corn
causes mortality and sublethal effects, such as growth inhibition,
densities of pollen, but was variable because of the typical
reaction by larvae to Bt proteins of feeding cessation followed by
extended periods of time before death (3). Growth inhibition, a
more sensitive measure of protein intoxication, could be de-
tected at 5–10 grains?cm2. Pollen from all other events, including
Mon810 and Bt11 corn hybrids as well as events not presently
Cry9C, and Cry1F proteins, respectively) did not demonstrate
lethal or sublethal effects, even at densities above 1,000 pollen
grains?cm2(3). These data were used to establish a no-
observable-effect-level for growth inhibition of larvae for event
176 pollen and for Bt11 and Mon810 pollen.
Five field bioassays were undertaken to determine the outcome
of exposure of larvae under field conditions on milkweed plants
growing or placed in the field. Field studies performed in Iowa,
Maryland, New York, and Ontario incorporating natural levels of
pollen from Bt corn plants (6) demonstrated no acute effects of
Bt11 and Mon810 corn pollen on survival or growth of monarch
larvae. However, impacts of event 176 pollen were observed. In
Iowa, reduced weight gain was noted for larvae exposed to event
176 pollen on milkweeds within cornfields at a density of 23 pollen
grains?cm2. Both survival and weight gain were affected in Mary-
land, where a series of assays using milkweed leaves collected from
plants in an event 176 cornfield were carried out over the pollen-
cornfield, pollen concentrations on milkweed leaves reached an
average of 67 and 161 grains?cm2. Survival of first instars was
to milkweed leaves collected from outside Bt cornfields. Weight
gain of survivors was reduced because of consumption of Bt pollen,
but only significantly so after exposure to pollen accumulated over
a 6-day period (6).
In a separate field trial in Maryland, effects on survival and
growth of first instar monarchs on leaves of milkweed within a
field of a sweet corn hybrid expressing Bt11 endotoxin were
evaluated and compared with the effects of residues after
applications of a pyrethroid insecticide. Survival of larvae that
fed on insecticide-treated milkweed leaves from within the
cornfield was low (0–10%). Survival also was influenced signif-
icantly (65–79%) by insecticide that drifted onto milkweeds
leaves 3 m outside the field. In contrast, survival of larvae
exposed to leaves taken from within both Bt and non-Bt corn
plots ranged from 80 to 93%, and there were no significant
differences in larval survival between these two plots (6).
Characterization of Exposure to Bt Corn Pollen. Exposure depends
on (i) phenological overlap between monarch populations and
corn anthesis, (ii) spatial overlap between milkweeds used by
monarchs and cornfields, and (iii) pollen densities encountered
on leaves of milkweed plants in and near cornfields.
Phenological Overlap. Pollen from corn plants within a particular
field is shed over a period of 1–2 weeks between mid-July and
Bt corn pollen on populations of the monarch butterfly.
Conceptual model of components of risk assessment of the impact of
www.pnas.org?cgi?doi?10.1073?pnas.211329998Sears et al.
mid-August during the season, whereas larvae develop over a
more prolonged period. Potential for exposure of susceptible
stages of monarch larvae to corn pollen depends on synchrony
of their development with pollen shed of corn plants. Locations
in four corn-growing regions were monitored for phenological
development of monarch populations and anthesis (4). These
locations were established in Iowa, Maryland, Minnesota?
Wisconsin, and Ontario. Overlap of the more susceptible stages
of monarchs, primarily first and second instars, with pollen shed
was considered for purposes of risk assessment.
The presence of susceptible larvae at the time of corn anthesis
varied considerably across the regions studied (4). In the more
northern locations (Minnesota?Wisconsin and Ontario), about
40 and 62% of the larvae overlapped with pollen shed, respec-
tively, whereas in areas further south (Iowa and Maryland),
about 15 and 20% of the larval stages overlapped, respectively.
Data from a computer simulation of monarch phenology and
corn development support the general observation that overlap
increases at higher latitudes across the Corn Belt (D. Calvin,
Department of Entomology, Pennsylvania State University,
University Park, PA, personal communication). Projected over-
lap from these simulations is likely overestimated because in the
model, 30-year average temperature data were used, whereas
specific fields at each of four locations.
Spatial Overlap. Density of milkweed stands in cornfields com-
pared with nonagricultural lands and data on the proportion of
the landscape in corn and nonagricultural lands provided a basis
on which to determine the proportion of the milkweed popula-
tion that was in cornfields (4). In all locations, densities were
higher in nonagricultural lands than in cornfields, but the range
of difference was considerable. In Minnesota?Wisconsin and
Iowa, the density of milkweed was approximately four to seven
in Ontario, the density was up to 115 times greater. In areas
where corn is more intensively cultivated, as in Iowa and
southern Minnesota?Wisconsin, less nonagricultural land exists,
and the overall proportion of milkweed on a landscape basis is
higher in cornfields and other crop lands than in nonagricultural
land. In regions of the corn-growing area where mixed habitats
are more common, such as in Maryland and Ontario, milkweeds
are more abundant in the nonagricultural landscape and provide
proportionately greater habitat than those in cornfields (4).
The likelihood of monarch larvae feeding on milkweed plants in
cornfields depends not only on what proportion of milkweeds are
in cornfields but also on the relative usage by monarchs of milk-
weeds in cornfields relative to milkweeds in other habitats. Obser-
vations from the four regions studied indicated that monarch
butterflies locate and lay eggs on milkweeds in corn despite the
canopy of the crop obscuring the milkweeds. Sampling done in
cornfields and in nonagricultural land in these areas suggests that
egg densities per plant are higher in corn in Iowa and Minnesota?
Wisconsin but are the same in cornfields and nonagricultural lands
in soybean fields than in nonagricultural areas.
Pollen Densities Encountered. Dispersal of corn pollen was de-
scribed by Raynor et al. (27), who demonstrated deposition of
pollen as much as 60 m from field edges. Because of the rapid
risk assessment concerns are focused on the concentration of
pollen on milkweed leaves within the cornfield and those leaves
found outside the field up to 5 m from the field edge. During the
period of pollen shed, samples of pollen were collected on sticky
trap surfaces and on milkweed leaves (5). Samples were taken at
various distances within and beyond the margins of cornfields to
estimate the concentration of pollen that could be encountered
by monarch larvae. Data from three locations, Iowa, Maryland,
and Ontario (5), demonstrated a 5-fold reduction in concentra-
tion of pollen from just within the edge of the cornfield to about
2–3 m distant. Within-field densities across the different studies
averaged between 65 and 425 pollen grains?cm2on milkweed
leaves at the peak of corn anthesis, with an average of 171
to determine the likelihood of encounter by first or second
instars of different concentrations of pollen within and outside
cornfields. These frequency distributions can be used to deter-
mine the probability of a larva encountering a toxic pollen dose.
Risk Characterization. To determine risks to monarch larvae asso-
ciated with Bt corn pollen, two components of greatest significance
are: (i) the frequency with which effective environmental concen-
trations exceed the thresholds for mortality or sublethal effects,
such as growth inhibition, of each Bt pollen type, and (ii) the
that are exposed to toxic levels of Bt pollen.
Probability of Toxicity. It is clear from both laboratory and
field-based studies (3, 6) that pollen from the dominant com-
mercial Bt corn hybrids (Mon810 and Bt11) does not express
Cry1Ab protein to a level that will impact monarch populations
to any significant degree. Hellmich et al. (3) suggested a con-
servative lowest-observable-effect-concentration (LOEC) be
established for these hybrids at 1,000 pollen grains?cm2of
milkweed leaf surface on the basis of a combined analysis of
laboratory bioassays exposing larvae to 1,000–1,600 of pollen
grains?cm2. Growth inhibition was evident for larvae exposed to
event 176 pollen at 5–10 grains?cm2, the lowest dose where
activity was noted by Hellmich et al. (3), therefore the effective
environmental concentrations for event 176 corn pollen will
frequently exceed this threshold in fields where it is planted.
Probabilities of toxicity for events 176, Bt11, and Mon810
pollen are depicted in Fig. 2 as a dose–effect relationship for
exposure of larvae to pollen plotted on log-probability scales
following methods accepted by the EPA (http:??www.epa.gov?
NCEA?ecorisk.htm). Growth inhibition of first instar monarchs
in response to increasing concentrations of event 176 pollen, as
reported by Hellmich et al. (3), is illustrated, with a no-
observable-effect-level at 5–10 pollen grains?cm2. In compari-
son, a hypothetical response curve for Bt11 and Mon810 pollen
is depicted by using the same slope parameter for the event 176
corn hybrids. These two response curves are compared with the average
cumulative proportion of pollen on milkweed leaves as recorded within Iowa
and Ontario cornfields during 1999–2000.
Percent growth inhibition at 96 h for monarch larvae exposed to
Sears et al. PNAS ?
October 9, 2001 ?
vol. 98 ?
no. 21 ?
a LOEC established, for sake of argument, as a range between
1,000 and 4,000 grains?cm2. Pollen deposition on milkweed
leaves during 1999–2000 (5) is represented on a separate scale
in a cumulative frequency occurrence curve.
It is apparent from Fig. 2 that significant overlap of commonly
encountered concentrations of pollen from event 176 hybrids with
doses necessary for growth inhibition (and mortality) would likely
occur within or near cornfields. For example, over 50% of a cohort
of first instars exposed to 10 pollen grains?cm2on milkweed leaves
would be expected to exhibit growth inhibition. When compared
with the pollen deposition curve, 90% of the samples of milkweed
leaves examined in our study of pollen deposition had a density at
or above this level. By contrast, the LOEC range for Bt11 and
Mon810 of 1,000–4,000 pollen grains?cm2would be encountered
by larvae in only 0.7–0.1% of natural in-field situations. Pollen
exceeded 1,600 grains?cm (3, 5). Even if the LOEC for Bt11 and
grains?cm2, 99.3% of encounters by monarch larvae of pollen
would be below this concentration (5).
To illustrate the potential impact of each pollen type on first
instar monarchs exposed to event 176, Bt11, and Mon810 pollen
within cornfields, a joint probability curve (Fig. 3) was con-
structed (http:??www.epa.gov?NCEA?ecorisk.htm; ref. 14). The
likelihood of exposure is plotted against the incidence of growth
inhibition as derived from data in Fig. 2. For event 176 pollen,
curve (AUC, 84%). In contrast, the impact for Bt11 corn pollen
(AUC, 0.1%) would be considered negligible.
Probability of Exposure.Milkweedsexistincornfieldsacrossmostof
use this resource as a host for their offspring during the period of
pollen shed (4). Quantification of the proportion of monarch
populations in each region potentially exposed to Bt corn pollen is
locations throughout the corn-producing area of North America,
illustrate the various factors influencing potential exposure to Bt
corn pollen (4). These data are insufficient to provide a definitive
estimate of exposure in most cases, but a bounding estimate is
possible. Wassenaar and Hobson (28), by using isotope analysis of
overwintering monarchs in Mexico, estimated that 50% of the
monarch population originates within all or part of 15 states and
Corn Belt (Table 1). More than 93% of North American corn is
western New York. By using USDA statistics, we estimate that
Adoption of Bt corn across this area encompassing 50% of the
monarch-breeding habitat was about 19% of the corn crop in
Distribution of milkweeds within and around cornfields is vari-
able across the Corn Belt. In Iowa and southern Minnesota,
milkweed in cornfields represents a large proportion of the total
represents a smaller proportion of the landscape, milkweed in
cornfields constitutes only a small proportion of its overall abun-
dance (4). These data provide a variable picture, but by using data
from Iowa, a suitable estimate of exposure to Bt corn pollen by
monarchs can be obtained. In addition to the fact that a significant
proportion of land in Iowa is devoted to corn production, a
milkweed plant in cornfields was 1.7 times more likely to receive a
monarch egg than a milkweed plant in nonagricultural land (4).
Even though milkweed densities are approximately seven times
higher in nonagricultural land than in cornfields, the proportion of
monarch larvae contributed by milkweed from within cornfields is
45 times greater than that of nonagricultural land. If all of the
landscape within Iowa that represents breeding habitat for mon-
archs is coupled with the higher per-plant egg densities on milk-
weeds within cornfields, the relative density of milkweeds in dif-
total habitat, 56% of monarchs in Iowa are estimated to originate
from within cornfields (4). This value probably differs considerably
from region to region, but we do not have comparable data for
other locations. From the data available (Table 1), we consider the
situation in Iowa to represent the upper end of the exposure scale.
Temporal overlap of pollen shed with the presence of sensitive
larvae in Iowa was 15% (4). An estimate of the probability of
exposure (Pe) to Bt corn pollen by larvae of the final monarch
generation arising within Iowa can be expressed as:
Pe? l ? o ? a ? 0.56 ? 0.15 ? 0.25 ? 0.021,
or 2.1% of the population, where Pe? probability of exposure,
l ? proportion of monarchs from corn, o ? overlap of pollen
shed with susceptible larval stages, and a ? adoption rate of Bt
corn. Because this represents an estimate of the potential
probability of exposure for other locations in the Corn Belt. We
do not have complete data for monarch and milkweed densities
across the Corn Belt and cannot assume that relative produc-
tiveness of crop and nonagricultural habitats is the same in other
states. For Iowa, the proportion of the monarch population
estimated to come from corn, 56%, was roughly similar to the
proportion of the breeding habitat in Iowa that is corn (42%;
Table 1). If we assume that this same relationship holds in other
areas, we can use the proportion of corn grown relative to the
total breeding habitat in other states (from Table 1) as an
estimate of relative monarch production.
Estimates of these three exposure factors and the estimated
contribution of each state and province to ?50% of the eastern
North American monarch population arising from the portion of
the Corn Belt, as indicated by Wassenaar and Hobson (28),
provides a broad view of potential exposure (Table 2). Our
estimates of overlap of the pollen-shed period in each location
monarch larvae to events 176 or Bt11 corn pollen with the percent of larvae
demonstrating inhibition of growth. The estimated area under the curves
provides a comparative measure of the impact that exposure to each type of
Joint probability curves that compare the likelihood of exposure of
www.pnas.org?cgi?doi?10.1073?pnas.211329998Sears et al.
with the presence of monarch larvae are based partly on the
projections of the simulation model described previously and
partly on our own observations. In this instance, our estimate for
the exposure of monarchs in the Corn Belt states and Ontario is
1.6%. Because monarchs in the Corn Belt represent 50% of the
total monarch population, the exposure for the entire monarch
population would be no greater than 0.8%.
The proportion of the population of monarchs in Iowa that
would be exposed to pollen levels that exceed the no-observable-
effect-level for each event (Pt) can be derived from data pre-
sented in Fig. 2. For event 176 pollen, monarch larvae would
likely encounter pollen densities equal to or exceeding the
LOEC in 90% of field situations during anthesis, whereas this
would be true in only 0.7% or less of field situations for Bt11 and
Mon810 pollen. Overall risk (R) is the combined probability of
exposure and toxic effect or:
R ? Pe? Pt.
If we assume that event 176 comprised 5% (a from Eq. 1 ? 0.05)
of planted corn acres in Iowa (or 20% of planted Bt corn acres
in 2000), an extreme upper bound estimate based on historical
marketing data, the risk of impact (R) to monarch populations
exposed to effects from event 176 pollen is:
R ? Pe? Pt? 0.0042 ? 0.9 ? 0.0038,
or 0.4% of the population. The LOEC of pollen for all other
events (Bt11 and Mon810 comprise the remaining 20% of total
area planted; a ? 0.20) equals or exceeds 0.7% of expected
pollen densities, thus the proportion of the monarch population
at risk of impact from effects of Cry proteins, other than event
176, in Bt cornfields in Iowa is:
R ? Pe? Pt? 0.0168 ? 0.007 ? 0.00012,
or 0.012% of the population. The combined risk estimate for
monarchs in Iowa is the sum of these two values, or 0.41%.
Following the same logic as above and assuming that (i)
adoption rate of Bt corn reached its maximum limit of 80%
(based on current refuge requirements) (a ? 0.80) in Iowa, and
(ii) pollen from current and future Bt corn events will pose a
hazard less than or equal to that established here for Bt11 and
Mon810, the proportion of the monarch population in Iowa that
would be at risk with market saturation is:
R ? Pe? Pt? 0.067 ? 0.007 ? 0.00047,
Table 1. Crop area for corn production in the Corn Belt states and provinces (2000) and the relationship of Bt corn to monarch
butterfly breeding range
Land area, ha (2000)
Corn as % of
crop and pasture
% Bt corn
of crop and
% crop and pasture area
w?in 50% monarch
% area of
Total186,685 101,01627,83526.517.8 4.459,3461.00
*Ref. 30; †, http:??usda.mannlib.cornell.edu?mor_start.html; ‡, http:??usda.mannlib.cornell.edu?reports?nassr?field?pcp-bba?acrg0600.pdf; §, Ref. 28.
Table 2. Parameter estimates for probability of exposure (Pe) of
monarch larvae to Bt corn pollen within cornfields in states and
provinces of the Corn Belt that constitute 50% of the eastern
North American monarch population
Parameter estimates for risk of
exposure in each state
Average (totals)0.2530.280.17 (1.020)(0.0161)
*l, proportion of monarchs from corn; o, overlap of pollen shed with suscep-
tible larval stages; a, adoption rate of Bt corn.
†Values marked with * were derived from field observations in 2000 (see
‡m represents the proportion land area of each state or province that consti-
tutes 50% of the breeding habitat of the eastern North American monarch
Sears et al.PNAS ?
October 9, 2001 ?
vol. 98 ?
no. 21 ?
or 0.05% of the Iowa population. If, instead, only event 176 Download full-text
hybrids were grown to the maximum extent in Iowa, 6.1% of the
monarch population would be at risk. By using this format and
data from Tables 1 and 2, risk of exposure and toxicity from Bt
corn can be applied to each of the states and provinces in which
monarch breeding and corn production overlap.
Previous reports (1, 8) indicating the hazard of Bt corn pollen to
monarch butterfly are inadequate to assess risk, because assigning
can be properly expressed and the likelihood of exposure is
estimated through appropriate observations. We have used a
comprehensive set of new data and a formalized approach to risk
assessment that integrates aspects of exposure to characterize the
risk posed to monarchs from Bt corn pollen. Characterization of
acute toxic effects alone indicates that the potential for hazard to
monarchs is currently restricted to event 176 hybrids, which express
Cry1Ab protein in pollen at a level sufficient to show measurable
effects. Event 176 hybrids have always had a minor presence in the
corn market and current plantings, which comprise ?2% of corn
acres, are rapidly declining.
Other events either express negligible Cry1Ab protein in corn
pollen (Mon810 and Bt11) or express Cry protein of significantly
Cry1Ac, Cry9c, and Cry1F proteins, respectively). These corn
hybrids have little or no effect on monarch populations, although
sublethal effects due to chronic exposure to Bt pollen over the
entire larval growth of monarchs has not been accounted for in
these studies. Should chronic effects he documented, the impact
on monarch populations will remain low or negligible, because
overall exposure of monarch larvae to Bt pollen is low.
portion of the monarch population that is potentially exposed to
toxic levels of Bt corn pollen is negligible and declining as planting
of event 176 hybrids is phased out through 2003. The exposure
toxicity portion (Pt) of this equation for the dominant corn hybrids
is negligible, therefore the impact of Bt corn on monarch popula-
tions should remain low.
Evidence supporting this risk conclusion has been collected
over a wide geographic area and under a variety of conditions in
both laboratory and field settings (3–6). Findings from studies
done in multiple locations were consistent, even though methods
differed from one study to another. This approach to risk
characterization is consistent with accepted risk assessment
procedures and shares many similarities with previous assess-
ments over a wide range of situations describing potential risk
associated with a described hazard. It is imperative that future
conclusions concerning the environmental or nontarget impacts
of transgenic crops be based on appropriate methods of inves-
tigation and sound risk-assessment procedures.
We thank Jeffrey Wolt, Keith Solomon, and Anthony Shelton for their
input and critical comments and suggestions during the development of
this paper. This research was supported by a pooled grant provided by
USDA–ARS and the Agricultural Biotechnology Stewardship Technical
Committee (ABSTC), and by funding from the Canadian Food Inspec-
tion Agency (CFIA), Environment Canada, and the Ontario Ministry of
Agriculture, Food and Rural Affairs, the Maryland Agricultural Exper-
iment Station, and the Leopold Center for Sustainable Agriculture,
Ames, IA. Members of ABSTC are Aventis CropScience USA LP, Dow
AgroSciences LLC, E. I. du Pont de Nemours and Company, Monsanto
Company, and Syngenta Seeds, Inc.
1. Losey, J. E., Rayor, L. S. & Carter, M. E. (1999) Nature (London) 399, 214.
2. U.S. Environmental Protection Agency (1995) Publ. No. EPA731-F-95–004
(U.S. Govt. Printing Office, Washington, DC).
3. Hellmich, R. L., Siegfried, B. D., Sears, M. K., Stanley-Horn, D. E., Daniels,
M. J., Mattila, H. R., Spencer, T., Bidne, K. G. & Lewis, L. (2001) Proc. Natl.
Dively, G., Olson, E., Pleasants, J. M., Lam, W.-K. F. & Hellmich, R. L. (2001)
Proc. Natl. Acad. Sci. USA 98, 11913–11918. (First Published September 14,
5. Pleasants, J. M., Hellmich, R. L., Dively, G., Sears, M. K., Stanley-Horn, D. E.,
Mattila, H. R., Foster, J. E., Clark, P. L. & Jones, G. D. (2001) Proc. Natl. Acad.
Sci. USA 98, 11919–11924. (First Published September 14, 2001; 10.1073?
6. Stanley-Horn, D. E., Dively, G. P., Hellmich, R. L., Mattila, H. R., Sears, M. K.,
Rose, R., Jesse, L. C. H., Losey, J. F., Obrycki, J. J. & Lewis, L. (2001) Proc.
Natl. Acad. Sci. USA 98, 11931–11936. (First Published September 14, 2001;
7. Hellmich, R. L. & Siegfried, B. D. (2001) in Genetically Modified Organisms in
Agriculture–Economics and Politics, ed. Nelson, G. C. (Academic, London), pp.
8. Jesse, L. C. H. & Obrycki, J. J. (2000) Oecologia 125, 241–248.
9. Wraight, C. L., Zangerl, A. R., Carroll, M. J. & Berenbaum, M. R. (2000) Proc.
Natl. Acad. Sci. USA 97, 7700–7703. (First Published June 6, 2000; 10.1073?
10. Giddings, J. M., Hall, L. W., Jr. & Solomon, K. R. (2000) Risk Anal. 2, 545–572.
E. E. (1999) Rev. Environ. Contam. Toxicol. 160, 1–129.
12. Solomon, K. R., Baker, D. B., Richards, P., Dixon, K. R., Klaine, S. J., La Point,
T. W., Kendall, R. J., Weisskopf, C. P., Giddings, J. M., Giesy, J. P., et al. (1996)
Environ. Toxicol. Chem. 15, 31–76.
13. Solomon, K. R., Giesey, J. P., Kendall, R. J., Best, L. B., Coats, J. R., Dixon,
K. R., Hooper, M. J., Kenaga, E. E. & McMurry, S. T. (2001) Environ. Toxicol.
Chem. 7, 497–632.
15. Klaine, S. J., Cobb, G. P., Dickerson, R. L., Dixon, K. R., Kendall, R. J., Smith
E. E. & Solomon, K. R. (1996) Environ. Toxicol. Chem. 15, 21–30.
16. Hall, L. W., Jr., Giddings, J. M., Solomon, K. R. & Balcomb, R. (1999) Crit.
Rev. Toxicol. 29, 367–437.
17. Kendall, R. J., Lacher, T., Jr., Bunck, E. C., Daniel, F. B., Driver, C., Glue,
G. E., Leighton, F., Stansley, W., Watanabe, P. G. & Whitworth, M. (1996)
Environ. Toxicol. Chem. 15, 4–20.
18. Kreig, A. & Langerbruch, G. A. (1981) in Microbial Control of Pests and Plant
Diseases, ed. Burges, H. D. (Academic, New York), pp. 837–896.
19. Peacock, J. W., Schweitzer, D. F., Dale, F., Carter, J. L. & Dubois, N. R. (1998)
Environ. Entomol. 27, 450–457.
20. Miller, J. C. (1990) Am. Entomol. 36, 135–139.
21. Johnson, K. S., Scriber, J. M., Nitao, J. K. & Smitley, D. R. (1995) Environ.
Entomol. 24, 288–297.
22. Hall, S. P., Sullivan, J. B. & Schweitzer, D. F. (1999) USDA Bull. No.
FHTET-98–16 (USDA, Washington, DC).
23. Malcolm, S. B., Cockrell, B. J. & Brower, L. P. (1993) in Biology and
Conservation of the Monarch Butterfly, eds. Malcolm S. B. & Brower L. P.
(Natural History Museum of Los Angeles County, Los Angeles, CA), pp.
24. National Research Council (1983) Risk Assessment in the Federal Government:
Managing the Process (Natl. Acad. Press, Washington, DC).
25. Koziel, M. G., Beland, G. L., Bowman, C., Carozzi, N. B., Crenshaw, R.,
Crossland, L., Dawson, J., Desai, N., Hill, M., Kadwell, S., et al. (1993)
Biotechnology 11, 194–200.
26. Christensen, A. H., Sharrock, R. A. & Quail, P. H. (1992) Plant Mol. Biol. 18,
27. Raynor, G. S., Ogden, E. C. & Hayes, J. V. (1972) Agron. J. 64, 420–427.
28. Wassenaar, L. I. & Hobson, K. A. (1998) Proc. Natl. Acad. Sci. USA 95,
29. USDA–National Agricultural Statistics Service (2000) Census of Agriculture
(USDA–NASS, Washington, DC), Vol. 1, Part 57.
30. USDA, Natural Resources and Conservation Service, & Iowa State University,
Statistics Laboratory (1997) Summary Report, National Resources Inventory
(USDA, Washington, DC).
www.pnas.org?cgi?doi?10.1073?pnas.211329998Sears et al.