    Environmental Protection Agency


                                       

White Paper on Corn Rootworm Resistance Monitoring for Bt Plant-Incorporated Protectants


                              September 27, 2013



				Biopesticide Pollution Prevention Division
                         Office of Pesticide Programs

EXECUTIVE SUMMARY

As of 2013, there are four Bt Plant-Incorporated Protectants (PIPs) registered for control of corn rootworm (CRW, Diabrotica spp.).  These PIPs include Cry3Bb1 (Monsanto Company), Cry34/35 (Dow AgroSciences and Pioneer Hi-Bred), mCry3A (Syngenta Seeds Inc.), and eCry3.1Ab (Syngenta Seeds Inc.).  In accordance with the terms of registration for each of these PIPs, registrants are required to implement a resistance monitoring program for CRW and, if resistance is documented, a remediation program.   

Resistance monitoring has two major components:  1) annual sampling of CRW populations from high resistance risk areas of the Corn Belt, and 2) investigations of populations from Bt fields with unexpectedly high levels of CRW damage.  Collected populations are tested for Bt susceptibility with bioassays using artificial diet or, in some cases, Bt corn plants.  CRW are considered resistant if the assays demonstrate a reduction in susceptibility to Bt (typically compared to a laboratory reference colony or historical data) that is heritable and can be linked to field-level economic damage.  Resistant populations are subject to remedial action plans intended to control and/or limit the spread of resistance.

BPPD has identified a number of scientific uncertainties with the CRW resistance monitoring program.  These uncertainties (detailed in this document) include questions regarding sampling protocols, bioassays to detect resistance, definitions of resistance for CRW (in the context of bioassay results), and remediation efforts.  Overall, BPPD is concerned that the monitoring program is reactive instead of proactive, such that resistance may evolve in the field at economically damaging levels before it is detected.  The timing of the bioassays to confirm potential resistance may take a year or more, potentially limiting timely ability to successfully implement remedial measures to preserve susceptibility.

Aspects of CRW biology present challenges for effective resistance monitoring.  Larvae feed on corn root systems below ground and cannot be directly observed.  By the time field damage becomes apparent, usually late in the summer, the CRW seasonal cycle may be complete and no adults may be available to collect for testing.  CRW undergo an obligate diapause period, meaning that populations collected during a growing season are not tested until the following spring.  CRW are also generally less susceptible to Bt toxins than other target pests (none of the CRW toxins are considered "high dose"), which complicates the field thresholds used to investigate unexpected damage. Further, the insect has also proven highly adaptable to multiple cultural and insecticidal control measures.  

RECOMMENDATIONS TO IMPROVE CRW MONITORING

A number of proposals to improve CRW resistance monitoring are discussed in this document.  Improvements regarding insect sampling, field-level screening for potentially resistant populations, the use of on-plant bioassays to detect resistance (instead of testing with artificial diet), definitions of resistance based on assay results, and remediation strategies are discussed below.  Specific recommendations are as follows:


1.      Sampling:
 Adopt a focused, non-random sampling approach for annual resistance monitoring in the Corn Belt, with an emphasis on areas with previous Bt failure or high risk factors for resistance;
 Conduct limited, non-random sampling can also be conducted in "fringe" areas of the Corn Belt (i.e., areas with few Bt corn failures and lower selective pressure for resistance);
 Use the Node Injury Scale (NIS, Oleson et al. 2005) in Bt corn fields with unexpected damage to identify CRW populations to investigate for potential resistance (NIS > 1.0 for single Bt toxins; NIS > 0.5 for pyramided Bt toxins);
 In response to unexpected damage that exceeds the NIS threshold, collect CRW from:
 within the problem site by using transect sampling through the damaged plot if adults are present;
 sentinel plots some distance away from the affected field to determine average resistance allele frequency in the population;
 sentinel plots nearby (with the same maturity as Bt fields) the following year if insects were absent from the damaged plot the prior year.

2.      Resistance Detection Assays:
 Replace artificial diet testing with on-plant assays as the primary diagnostic resistance detection tool for CRW.  Two such assays, Nowatzki et al. (2008) and Gassmann et al. (2011), have been developed to date; 
 Perform at least two susceptibility measures as part of the on-plant assay, such as percent mortality, larval weight, body surface area, or instar development.

3.      Resistance confirmation
 Base definitions for resistance on an on-plant assay approach that incorporates two independent measures of susceptibility  -  e.g., the putative field population tested on Bt vs. non-Bt corn and the field population vs. a susceptible laboratory colony tested on Bt corn;
 Use scientifically sound statistical criteria (or resistance ratios) for resistance determinations. The approach should be sensitive enough to reduce the likelihood of false negatives (i.e., not detecting a resistant population) while also minimizing false positives.

4.      Remedial action
 Promote the use of crop rotation and alternate Bt corn modes of action as the preferred alternate CRW control strategies for resistance mitigation;
 Use of insecticides in the remediation strategy is recommended for short-term and local population suppression only;
 Develop methods to identify remedial action areas (e.g., by using resistance allele gradients or other tools);
 Use theoretical models to design species and toxin-specific remedial action plans (ideally before resistance has evolved), including the following components: 
 An assessment of remediation strategies for local and widespread resistance scenarios;
 Simulations to evaluate the percentage of CRW dispersal that would make containment unsuccessful (research may be needed to identify the proportion of long distance movement to be modeled);

DOCUMENT ORGANIZATION

This white paper is organized into a background section followed by six sections describing the corn rootworm (CRW) resistance monitoring program.  The first section (Problem Formulation) broadly identifies BPPD's concerns and scientific uncertainties with CRW resistance monitoring.  The second section describes sampling procedures for both annual resistance monitoring and in cases of unexpected damage to Bt fields.  Bioassays to detect resistance are detailed in the third section.  The interpretation of bioassay data in the context of defining resistance is the subject of the fourth section.  Resistance monitoring for northern corn rootworm is addressed in the fifth section.  The last section covers remediation (steps to be taken in the event of resistance).  A list of references is also provided at the end of this document.

BACKGROUND

Corn rootworm (CRW) is among the most serious economic insect pests of corn in the United States (Levine & Oloumi-Sadeghi, 1991). Northern corn rootworm (D. barberi, NCRW) is found in the Midwest of the U.S and has a localized distribution; Western corn rootworm (Diabrotica virgifera virgifera, WCRW) is the more serious pest of the two and is more widely distributed (from Mexico to Canada). Together, they are the most prevalent Diabrotica pests in the U.S., and WCRW has been commonly referred to as a "billion dollar pest."  In this document, the use of "CRW" primarily refers to WCRW, given that it is the dominant rootworm species in the Corn Belt and the main target pest of Bt corn.  

Bt Corn PIPs Registered for Control of CRW

At the time of this review, four Bacillus thuringiensis (Bt) toxins have been registered for control of CRW in corn plant-incorporated protectants (PIPs) (see Table 1).  The first of these, approved by the Agency in 2003, was the Cry3Bb1 protein in Monsanto Company's MON 863 (YieldGard Rootworm, EPA Reg. No. 524-528).  This was followed in 2005 by the registration of MON 88017 corn (Yieldgard VT Rootworm, EPA Reg. No. 524-551). The Cry3Bb1 toxin expressed in MON 88017 is equivalent to that in MON 863 and also a variant of the wild-type Cry3Bb1 protein from Bt subspecies kumamotoensis. When compared by amino acid sequencing, the Cry3Bb1 protein expressed in MON 88017 has been reported to be 99.8% similar to the Cry3Bb1 protein expressed in MON 863.  The primary difference between the hybrids is that MON 88017 also expresses a gene for resistance to glyphosate (Roundup) based herbicides.

Also in 2005, EPA approved registration of Bt corn products containing the Cry34/35Ab1 binary toxin, developed by Dow AgroSciences and Pioneer Hi-Bred as part of the "Herculex Rootworm" product lines (EPA Reg. Nos. 29964-4 and 68467-5).  The third and fourth CRW Bt PIPs were registered by Syngenta Seeds Inc. and include mCry3A (registered in 2006 as Event MIR604; EPA Reg. No. 67979-5) and eCry3.1Ab (registered in 2012 as Event 5307; EPA Reg. No. 67979-22).

Table 1.  Registered Bt Corn PIPs for CRW Control

                                     Toxin
                                     Event
                                  Trade Name
                                  Registrant
                           Year Initially Registered
                                    Cry3Bb1
                                    MON 863
                                   MON 88017
                             Yieldgard VT Rootworm
                                   Monsanto
                                     2003
                                  Cry34/35Ab1
                                    59122-7
                               Herculex Rootworm
                       Dow Agrosciences, Pioneer Hi-Bred
                                     2005
                                    mCry3A
                                    MIR 604
                                  Agrisure RW
                                Syngenta Seeds
                                     2006
                                   eCry3.1Ab
                                     5307
                               Agrisure Duracade
                                Syngenta Seeds
                                     2012


Subsequent to the initial single toxin registrations, these CRW PIPs have been approved in a number of other configurations, including pyramids with two or more CRW-active toxins (e.g., SmartStax), varieties stacked with lepidopteran traits (e.g., Herculex Xtra), and seed blends (e.g., Optimum AcreMax).

Corn Rootworm Resistance Monitoring

Through the terms of registration, Bt corn registrants are required to implement a CRW resistance monitoring program for their traits.  A primary goal of resistance monitoring is to detect shifts in the frequency of resistance genes (i.e., susceptibility changes) before the onset of resistance leads to widespread Bt crop failure (see BPPD 2010a).

Monitoring for CRW resistance has consisted of two main components: 1) investigation of unexpected field damage reports from growers, extension agents, consultants, or company agronomists, and 2) monitoring for changes in susceptibility through targeted population sampling and testing.  Unexpected damage reports may reveal the occurrence of localized resistance (or hot spots) before the effects can spread.  Resistance monitoring through targeted field sampling can reveal changes in susceptibility of geographically representative populations. 

The main elements required for CRW resistance monitoring are generally consistent between registrants and have included the following:
 Investigations of grower, extension specialist or consultant reports of less than expected results or CRW control failures based on root damage standards and other field criteria.
 CRW sampling focused in areas in which there is the highest risk of resistance development (coordinated for all Bt corn registrants by the Agricultural Biotechnology Stewardship Technical Committee).
 Bioassays to determine the susceptibility of each sampled population.  Baseline susceptibility studies for each registered trait are used to establish a reference for monitoring of future CRW populations.
 Development and validation of an appropriate discriminating or diagnostic dose assay as a resistance detection tool. 
 Definitions of resistance (based on bioassay results) to trigger remedial action plans.

Each of these components is discussed in greater detail in this review and in Biopesticide Registration Action Documents ("BRADs") for each registered trait (see BPPD 2010a, b, c).  It should be noted that the monitoring program for eCry3.1Ab trait (registered in 2012) is still under development and will be implemented when the product is commercialized.  

In 2010, EPA reassessed and extended the registrations for all Bt corn PIPs.  As part of the extensions, additional monitoring terms were implemented for CRW.  A number of these were in response to concerns raised by BPPD in its assessments of the monitoring plans and data submitted after registration (see discussions in BPPD 2010a, b, c).  These terms include the following (paraphrased from registration notices in BPPD 2010d):

 Development of functional diagnostic assay to detect potentially resistant CRW populations.  As part of the efforts, the registrants were required to investigate the feasibility of using the Sublethal Seedling Assay (Nowatzki et al. 2008).
 Development and implementation of a monitoring plan for northern corn rootworm (NCRW).  This plan must include insect sampling and testing of NCRW susceptibility to Cry34/35.
 Revised CRW damage guidelines for Cry34/35 corn (for reports of unexpected pest damage in the field) incorporating BPPD's comments from the June 30, 2010 review (see BPPD 2010a, b, c). 
 Investigation of grower, extension specialist or consultant reports of unexpected damage or control failures for corn rootworm.

Information submitted to fulfill the terms has been reviewed by BPPD (see BPPD 2011, 2012, 2013a, 2013b).  Specific scientific issues related to the resistance monitoring programs are discussed in those reviews and in this document.


I.   PROBLEM FORMULATION

Several recent studies by research scientists have documented cases of field-evolved corn rootworm (CRW) resistance to Bt corn.  The first of these studies, Gassmann et al. (2011), demonstrated that CRW collected from damaged Cry3Bb1 fields in Iowa during 2009 were significantly less susceptible to the toxin than CRW from undamaged or non-Bt fields. Additional cases of Cry3Bb1 damage characterized as field resistance have since been documented from other locations in Iowa and Illinois (Gassmann et al. 2012a, b, Gray 2012). These reports did not meet EPA's regulatory definition for resistance because the current definitions of resistance employ different investigative techniques.  Nonetheless, BPPD concurred with the scientific findings of the researchers that the results showed CRW resistance had evolved in some parts of the Corn Belt (BPPD 2012).  In addition to these resistance studies, preliminary data presented by three research teams at the 2012 Entomological Society of America meeting suggests that cross resistance between Cry3Bb1 and mCry3A may occur in resistant CRW populations (Gassmann 2012, Meinke et al. 2012, Zukoff and Hibbard 2012).  Further research will be needed, however, to fully assess potential cross resistance and implications for insect resistance management (IRM).

CRW resistance threatens one of the key benefits of Bt corn:  decreased chemical insecticide use (BPPD 2010a).  Bt fields with resistant populations are likely to be treated with pesticides to mitigate resistance and limit economic damage in future seasons.  As Bt corn efficacy becomes less certain, growers even in unaffected areas may seek prophylactic insurance by combining insecticide use with Bt corn planting.  In at least one state in the Corn Belt (Illinois), such practices were expected to increase during 2013 (Gray 2013).  Resistance to one Bt toxin can also put other registered Bt toxins at risk of failure, either through cross resistance or selection pressure, if multiple toxins are in a pyramided PIP with a small refuge (i.e., 5%).

In response to these concerns, BPPD is reviewing the CRW monitoring and remediation programs, identifying the scientific uncertainties associated with CRW resistance, and proposing improvements based on current available information.  Previous BPPD reviews (2011, 2012, 2013a, b) have also addressed CRW resistance monitoring; some of the scientific concerns and issues raised in those reviews are described in the sections below.  

a. Non-High Dose Toxins

The three currently commercialized Bt Cry toxins for CRW control (described in the Background section of this document) are not considered "high dose" as defined by the 1998 Science Advisory Panel (SAP, 1998).  EPA's refuge-based Insect Resistance Management (IRM) approach was developed for high-dose lepidopteran PIPs but was nonetheless applied to these CRW PIPs, largely to maintain consistency with existing IRM programs for lepidopteran-targeted Bt corn. When a toxin is high dose, the functional dominance is reduced and the resistance allele becomes recessive in nature. Conversely, direct experimentation with Diamondback moths (Plutella xylostella) and tobacco budworm (Heliothis virescens) demonstrated that resistance evolved fastest when the dominance of the resistance allele was increased by lowering the dose of toxin (Tabashnik et al., 2004). The greater risk of evolving resistance to Bt crops can be expected when toxins produced by the PIP are less than high dose. It is, therefore, not unexpected that first suspected resistance reports emerged in 2009 (Gassmann et al. 2011), approximately six years after the commercialization of Cry3Bb1.

Lack of high dose underscores the need for proactive resistance monitoring but also presents a number of challenges.  Field damage can be expected for lower dose crops, even if populations are susceptible to toxin.  This can complicate scouting and evaluations of unexpected pest damage (a key element of the CRW monitoring program).  Many fields are prophylactically treated with conventional insecticides to ensure crop viability (e.g., Gray 2013), which can be expected to mask observable damage on Bt corn from resistant populations.

b. BPPD Concerns with the Present Resistance Monitoring Program for CRW

The current resistance monitoring requirements for CRW are described in the Background section and in more detail in sections II through VI of this paper.  BPPD's reviews of submitted CRW resistance monitoring strategies and data have identified a number of scientific uncertainties with the approaches used to date (see BPPD 2010a, 2010b, 2010c, 2011, 2012, and 2013).  These issues are summarized below but are discussed in detail in subsequent sections.  Overall, current resistance monitoring plans are reactive rather than proactive.  BPPD believes the program has limited ability to detect shifts in the frequency of resistance genes (i.e., susceptibility changes) before the onset of resistance could lead to widespread crop failure.

Each year, approximately 12-15 WCRW populations are sampled from states in the Corn Belt (typically Illinois, Iowa, and Nebraska).  The locations are selected randomly, meaning that populations are not tracked over time for shifts in susceptibility.  Rather, the susceptibility of field populations is compared to laboratory reference colonies and the results from field populations tested in previous years.  Testing has been done largely with artificial diet bioassays, though these assays do not appear to be sufficient to reliably detect resistant CRW populations or function as "diagnostic" concentrations (to distinguish susceptible and potentially resistant individuals).  Results from the diet bioassays are often highly variable, such that comparisons to susceptible laboratory colonies or historic baselines are difficult to evaluate.
  
A second component of the monitoring program has been the assessment of reports of CRW unexpected damage to Bt fields.  Registrants evaluate these fields based on a set of criteria that, in addition to environmental and agronomic factors, includes Node Injury Score (NIS) triggers to determine whether CRW should be sampled for further investigation.  The use of NIS presents a number of challenges:  the injury-based threshold ideally should reflect decreased susceptibility of the population indicative of putative resistance, while avoiding false negatives (potential resistance) or false positives (unnecessary utilization of industry and Agency resources). The non-high dose nature of CRW PIPs is one factor, for example, that may confound NIS results; some damage should be expected with these Bt PIPs. Favorable environmental conditions can also lead to unusually high population numbers, which will lead to unexpected damage although the population may not be resistant. 

If field damage triggers are met, registrants attempt to sample CRW for further investigation with bioassays.  BPPD has recommended (BPPD 2010f) that sampling occur in the damaged field, given that adult CRW engage in predominantly intra-field dispersal rather than frequent long distance dispersal (Spencer et al. 2003, Nowatzki et al. 2003a and 2003b).  There has been some concern, however, that sampling within Bt fields can result in a bias in which less susceptible biotypes may be more likely to be collected (given that larvae are continually exposed to Bt).

Timing has been an additional concern with monitoring, particularly in regards to the ability to detect and mitigate potential resistance events.  Western and northern corn rootworm undergo an obligate (egg life stage) diapause in the winter, with some northern corn rootworm known to engage in an extended diapause through the following growing season and winter (areas in MN, IA, and SD).  Diapause complicates resistance monitoring by delaying bioassays; larvae collected in late summer are typically tested the following spring.  As such, any resistance determination will likely take at least one year (and possibly up to two) from the initial population sampling.  A protracted detection process may ultimately limit effective mitigation (centered on population containment) with a remedial action plan.

Current regulatory definitions of resistance (as defined in the terms of registration) establish criteria for "suspected" and "confirmed" resistance among tested populations.  Suspected resistance is based on damage triggers from Bt fields with unexpected CRW injury.  Confirmed resistance is determined with a standard diet bioassay and LC50 comparisons to historical baselines.  The resistance trait must also be shown to be heritable and confer the ability to inflict economically significant damage on Bt corn.  In practice, however, the use of diet testing with current laboratory methods to appears to be unworkable as a resistance detection tool.  Responses to toxin-incorporated diets are highly variable and a functional "diagnostic" concentration (capable of detecting resistant individuals or populations) has not been developed for any of the currently registered toxins.

Several research teams have developed corn-based ("on-plant") assays that offer potential as diagnostic CRW screening tools (Nowatzki et al. 2008, Gassmann et al. 2011).  These assays involve exposure of CRW larvae to corn plants under greenhouse conditions, which are then assessed for survival and/or sublethal effects.  Based on the development of these assays, registrants of CRW traits have been required to investigate the feasibility of using them for resistance detection.

Requirements for remedial action are generic but mandate that registrants work with EPA to develop a long-term plan.  Growers are encouraged to use different CRW control measures, and sales of the trait may be restricted in the affected region.  Recently, Monsanto has responded to reports of Cry3Bb1 resistance by working with growers to implement Best Management Practices (BMPs) based on alternate CRW management tactics (e.g., soybean rotation, conventional insecticides, and pyramided Bt corn). Generally, the goal of remedial action is to contain resistant populations to preserve the toxin in other areas where it is still effective.  Such an area-wide approach may not be practical with CRW, for which resistance may evolve on a field-by-field basis and long distance dispersal is limited (relative to lepidopteran corn insects).  On the other hand, a lengthy resistance confirmation process may challenge the ability to respond quickly to localized cases of Bt field resistance. 

II. CORN ROOTWORM SAMPLING

a. Annual Sampling of CRW Populations in the Corn Belt

As a condition of registration, registrants of Bt corn PIPs are required to conduct annual sampling of CRW populations from locations in the Corn Belt to monitor for resistance and/or trends in increased tolerance (see BPPD 2010d).  These collections have been coordinated by the Agricultural Biotechnology Stewardship Technical Committee (ABSTC) for all Bt corn registrants and are obtained from three defined regions covering different CRW biotypes. The regions also represent areas of high Bt corn adoption and CRW pressure. Region 1 (rotation resistant variety) includes Indiana and eastern Illinois; Region 2 (wild type) consists of western Illinois, Iowa, and Missouri; Region 3 (organophosphate resistant variant) encompasses Nebraska and Kansas. ABSTC attempts to collect multiple populations in each region, though some collections are not successful because insufficient offspring are produced for testing.  Typically, 12-15 total populations are collected each year, mostly from Iowa, Illinois and Nebraska.  A minimum of 2,000 adults is targeted for collection to establish each population for susceptibility testing. These adults are sampled from non-Bt hosts (e.g., refuge corn) during the growing season and are moved to a laboratory facility for rearing.  Egg progeny from the collected adults are maintained under diapause conditions until testing the following year.

While collections are obtained from random locations within these regions, the sampling focus has been in areas with historically large Bt corn adoptions and where the risk of resistance evolution is presumably greatest. Along these lines, BPPD has recommended that additional states with reports of unexpected damage be included in the sampling scheme; these consist of Wisconsin, South Dakota, Colorado, and Minnesota (see BPPD 2011, 2012).

Associated Uncertainties  -  Annual Sampling

The effectiveness of the monitoring program to proactively detect resistance is largely dependent on how population sampling is conducted.  To date, insect samples have been collected in a random fashion (within the three collection regions) from non-Bt hosts (e.g., refuge corn, pumpkin).  This type of sampling is used to determine whether a change has occurred in the average susceptibility of corn rootworm populations in different parts of the Corn Belt. Because the populations are selected randomly, however, this approach does not track susceptibility of individual populations (or populations within discrete geographic locations) over time.  More recently, BPPD has recommended switching to an intense and focused sampling approach in light of reports of Bt corn failures and resistance in the Corn Belt (see BPPD 2012). This shift was proposed to identify resistant hot spots as early as possible, ideally before a broad geographical spread of resistance genes can occur. The primary objectives of this change are to actively target areas of concern, determine annual sampling locations based on high risk factors, and identify any resistance problems (Roush & Miller 1986). BPPD has recommended the use of high risk (causal) factors of resistance to determine annual sampling locations. These considerations should include areas in close proximity to reported field failures/unexpectedly high rootworm damage, fields with a history of corn-on-corn planting (i.e., lack of crop rotation with a non-CRW host plant), repeated (e.g., prophylactic) use of the same Bt toxin, and/or high non-compliance with refuge requirements.

On the other hand, a totally focused sampling approach may be as limited as a completely random approach.  For example, it may be informative to compare targeted samples to a limited number of random samples from "fringe areas," i.e., areas where selection pressure is lower and where Bt corn adoption is not as prevalent. These fringe areas could serve as additional (quasi) baseline data. Assuming that long distance dispersal is limited in corn rootworm (excluding transportation by weather systems), this added random sampling approach would provide a point of comparison between the higher and lower Bt selection environments for corn rootworm and supplement the baseline data established in the early years of Bt commercialization. It may be appropriate to focus the majority of insect collections on areas where unexpected Bt damage is reported and high "risk factors" (as described above) are prevalent so that early resistance detection for non-high dose CRW toxins has a greater probability of success. 

The number of populations to be collected within each sampling region is another uncertainty.  With the present monitoring program, 12-15 populations are collected across all three regions.  It is unclear whether this number of populations is sufficient to adequately assess the susceptibility of CRW to the registered Bt toxins.  Sampling targets (total populations and number of individuals per population) could be based on a presumed resistance allele frequency for CRW.  To illustrate, if the phenotypic frequency of resistance is one in 1,000 (0.001), then more than 3,000 individuals must be sampled to have a 95% probability of detecting one resistant individual (Roush & Miller 1986).  For CRW and the different toxins, the frequency of the resistance alleles is unknown but is expected to be higher than 0.001 given that the currently-registered Bt PIPs have less than high dose toxin expression and the PIPs have been commercialized for some time.  If so, it may be possible to sample fewer individuals per population while maintaining a high probability of detecting resistant genotypes.  

The generally low susceptibility of CRW to Bt toxins (as evidenced by the lack of high dose expression) presents another challenge to proactive CRW monitoring.  A goal of annual monitoring has been to detect shifts in Bt susceptibility before field-level resistance results in crop damage (BPPD 2010e).  There is uncertainty, however, whether the bioassays for CRW (see Section III) are sensitive enough to be able to distinguish such shifts in susceptibility with any degree of precision.  Rather, it may be more likely that reduced susceptibility manifests as field damage (see part b below).

b. CRW Collections in Response to Unexpected Bt Damage

In addition to the annual sampling in the Corn Belt, CRW are collected in response to reports of damage in Bt fields.  Bt corn registrants are required to investigate all reports of unexpected product performance from growers, extension specialists or consultants (BPPD 2010d).  These investigations, often termed "performance inquiries," involve a number of procedural steps to rule out other potential causes of corn injury (the specific steps vary between registrations):  1) the damaged corn is confirmed as Bt (and not non-Bt or refuge corn), 2) Bt protein expression is at expected levels, 3) the damage is determined to be from CRW (and not a species with no susceptibility to Bt), 4) climatic, agronomic or other factors are eliminated as a factor in the damage.  

If CRW cause the Bt corn damage and the level of injury exceeds a root damage trigger based on the Iowa State Nodal Injury Scale (NIS; Oleson et al. 2005), then registrants are required to collect insects from the area.  For single Bt toxin products, the NIS trigger is 1.0 (i.e., one root node or the equivalent of an entire node, removed to within approximately 1(1/2) inches of the stalk); for pyramided products (expressing two or more CRW toxins), the trigger is 0.5 (i.e., (1/2) node or equivalent).  The NIS threshold is higher for single toxin products because they are expected to incur more damage under normal field conditions than pyramids with multiple modes of action to control CRW. If field damage triggers are exceeded, insect collections are to be made as soon as possible for testing to determine potential resistance (see section III for a discussion of the bioassays used to assess resistance).

Bt corn registrants have proposed different tactics at different points for sampling CRW in response to damaged fields.  CRW may be collected "in the immediate vicinity of the affected field" (Dow/Pioneer, MRID 484307-01 as discussed in BPPD 2013a) or as far away as 1-2 miles away from the field with reported damage (Monsanto, MRID 484368-01 as discussed in BPPD 2011). [Monsanto subsequently amended their Cry3Bb1 registrations in 2013 to specify that collections will occur in the damaged field and bordering fields as necessary.]  One uncertainty affecting the choice of sampling location is whether the resistance allele frequency within a damaged Bt field may be biased higher than in surrounding areas and therefore would not be representative of the overall CRW population in the vicinity (discussion at the January 2013 USDA NCCC-46 meeting, New Orleans, LA).  BPPD has concluded that adult collections should occur directly in the damaged Bt fields to determine whether resistant individuals were the cause of the Bt corn failure (BPPD 2011).  BPPD has further proposed that transect sampling be used through the damaged field site for collecting root node injury data. Adult insects should ideally be collected near the damaged Bt plants because of the limited adult dispersal. The objective of this type of data collection is to determine whether resistant individuals were responsible for the Bt failure. Collections should also be made some distance away from the failed field(s) to determine the average resistance allele frequency in the population. For this purpose, sentinel plots could be placed in the vicinity of failed Bt fields. Such plots would allow continued monitoring of resistant populations and help determine the resistance allele frequency in the population, as well as provide a source of insects in these areas the following year should best management practices be implemented by growers. 

The CRW sampling protocols typically do not describe the numbers of individuals to be collected from damaged fields.  [As amended in 2013, Monsanto's Cry3Bb1 registrations specify that at least 250, ideally 500, adults will be collected from the damaged field.]  While a robust sample is necessary to assess potential resistance, damage reports can occur later in the season (after root damage has become apparent), limiting opportunities to collect adults. In such situations, it may be necessary to analyze a smaller sample size to evaluate potential resistance, provided that appropriate statistical power can be obtained. 

If samples are obtained from a site the following year because insufficient adults were present at the site of reported crop damage during the initial investigation, these collections should occur in the same vicinity and irrespective of the level of damage in the Bt corn field. The planting of sentinel plots (some distance away that is reflective of the typical adult dispersal) would assure that a sufficient amount of insects will be available for collections. These plots, however, would need to have the same maturity as the surrounding corn in the area so as not to function as a sink (such as late planted corn). In most cases, BPPD expects that growers will implement added CRW control measures on damaged fields, including soil insecticide the following season to prophylactically treat CRW. These measures may lower population densities and reduce root damage (possibly below "unexpected damage" thresholds) despite the continued presence of resistance alleles in the population.  Further, the resistance allele frequency is unlikely affected by management practices (unless the problematic population is extirpated) and may persist in future generations. 

Associated Uncertainties  -  Sampling in Response to Unexpected Damage

The use of root node damage triggers (NIS scores) as a screening tool and/or potential indicator of resistance is discussed in Section IV of this document.

As discussed above, there is some uncertainty regarding where samples should be taken to investigate damaged fields.  It has been suggested that targeted adult sampling directly within failed Bt fields will bias susceptibility assays, given that larvae are continually exposed to the Bt toxin.  Insects in these Bt fields may still be "susceptible" but may have greater fitness (measured as percent survival) than insects from neighboring non-Bt fields.  This could potentially lead to false positives (calling a population resistant when in fact it is not).  BPPD does not dispute that insects sampled from Bt fields may show less susceptibility in assays; the purpose of collecting in these Bt fields, however, is to evaluate whether resistance has caused the Bt failure. Testing insects collected outside of the damaged field perimeter may not reveal much about the susceptibility of insects at the site of concern (unless resistance is area wide).  Since CRW resistance may evolve due to local management practices, BPPD asserts that adult insect collections for bioassays should be obtained from damaged fields, preferably close to the damaged site if collections occur the same year. Furthermore, BPPD believes that any observed effect (i.e., reduced susceptibility of CRW sampled from Bt fields) is more likely to be the result of a selective process for Bt resistance. All Bt corn products for CRW are non-high dose, meaning that insects carrying resistance genes (heterozygous genotypes or individuals with incomplete resistance/multiple resistance genes) will have a fitness advantage over susceptible ones. 

Some registrants have expressed concern with transect sampling through the damaged Bt site and have instead proposed a random sampling approach in the entire Bt field (for root node injury and adult collections).  With this approach, there is uncertainty in regards to obtaining root node injury samples that are used to reach `unexpected damage' and `suspected resistance' conclusions. Rootworms do not exhibit extensive average daily movement in a corn field (Spencer et al. 2009; Nowatzki et al. 2003a, b) and putative resistant females likely lay their eggs in nearby clusters of the field rather than throughout the entire field. Therefore, collecting root injury information across the field would reduce the overall average root injury results and could increase the likelihood for false negatives for "unexpected damage."

III. RESISTANCE DETECTION ASSAYS

a. Diet Bioassays

Historically, insect resistance monitoring for conventional pesticides has been conducted with diet bioassays using dose/response curves to make comparisons between LD/LC50s of reference (susceptible) strains and field-collected populations. Roush and Miller (1986) concluded that this type of comparison was adequate if resistance allele frequencies in a population were high but that it was not an effective method for early resistance detection. They proposed that a diagnostic assay approach, in which susceptible individuals are killed but resistant individuals survive, would be more efficient when monitoring for resistance.  However, the authors also noted that even with diagnostic tests, the required sample size needed to detect a resistance allele frequency of 0.01 could be very large. 

Both susceptibility (LC and EC) comparisons and diagnostic assays have been incorporated into the currently mandated IRM program for Bt corn PIPs to identify potentially resistant populations.   For Lepidoptera pests of corn (e.g., ECB, SWCB), diagnostic concentrations have been established (see discussion in BPPD 2010e).  On the other hand, for corn rootworm, no diagnostic concentrations are available to identify potentially resistant from susceptible individuals. BPPD has identified the lack of diagnostic concentrations for CRW as one of the shortcomings of the monitoring program and a primary reason why current diet bioassays are unlikely to proactively detect resistance (BPPD 2012). Presently, registrants analyze corn rootworm diet assay results by comparing mean LC/EC50s of field populations to control colonies or the 95% confidence interval of the historical baseline data. Because of the inherent variability with LC-measures derived with diet bioassays, registrants and EPA have relied more on EC50 data (calculated using development and body weight data). Some registrants have complemented these data with mean body weight results for survivors at the highest Bt toxin concentration. Although EC50 data were less variable, BPPD concluded that they were still not sensitive enough to make early resistance conclusions. Tables 2 - 4 show the historical EC/LC ranges and/or percent mortality, and average mass of survivors observed at the highest concentration tested with Cry3Bb1, Cry34/35, and mCry3A. 

Table 2.  Historical EC/LC50 values and body weight of surviving corn rootworm at the highest concentration of Cry3Bb1 tested (annual monitoring program and problem inquiries) (Table generated from data provided by Monsanto).

Year collected
                               Problem Inquiries
                      Annual Monitoring (Random Sampling)

                                    EC50[1]
                                   (95% C.I.)
                                  (ug/cm[2])
                                   EC50[1] 
                                  (95% C.I.)
                                  (ug/cm[2])
                                % Mortality[1] 
                                (170 μg/cm[2])
                         Average mass of survivors[1]
                                (170 μg/cm[2])
2007 
                                  12.68-38.38
                                 (10.73-60.06)
                                  14.20-33.46
                                  (8.2-40.8)
                                  36.11-91.67
                                   0.08-0.12
2008 
                                  7.73-44.50
                                 (4.57-93.50)
                                   7.3-30.4
                                  (2.1-43.1)
                                  36.11-97.97
                                   0.00-0.04
2009 
                               No PI collections
                                  13.9-63.61
                                  (6.8-113.0)
                                Not applicable
                                Not applicable
2010 
                                  46.55-82.53
                                (34.47-135.72)
                                 19.29-119.60
                                 (25.6-153.7)
                                  38.89-58.33
                                   0.10-0.16
2010 Field control on non-Bt diet
                                       
                                Not applicable
                                       
                                Not applicable
                                   13.89[4]
                                    0.25[4]
2011
                             3.3 to >341.6[2,3]
                                1.08-151.99[2]
                                 (0.10-201.97)
                                 170 μg/cm[2]
                                341.6 μg/cm[2]
                                 170 μg/cm[2]
                                341.6 μg/cm[2]

                                       
                                       
                                     19-58
                                     15-64
                                   0.13-0.37
                                   0.12-0.34
2011 Monsanto Lab Colony on Bt Diet
                                     10.23
                                  (3.1-16.7)
                                     90[5]
                                     89[5]
                                    0.12[5]
                                    0.11[5]
[1] Susceptibility data are presented as the range of results obtained from populations tested during that year.
2 The highest concentration tested in 2011 was 341.6 ug of Cry3Bb1/cm[2] 
[3] EC50s of some problem inquiries exceeded the highest concentration tested and could not be measured
[4] Weight and mortality of susceptible lab colony on isoline corn diet
[5] Weight and mortality of susceptible lab colony on Cry3Bb1 corn diet

Cry34/35 and mCry3A

Diet bioassays conducted with Cry34/35 and mCry3A have consisted of dose response tests to calculate EC (growth inhibition) and LC (mortality) values for sampled populations.  Field-collected populations are compared with susceptible laboratory colonies and the historical results for the toxin.  No testing with diagnostic concentrations has been conducted with either Cry34/35 or mCry3A.

Results from the Cry34/35 and mCry3A monitoring assays are summarized in Tables 3 and 4, respectively.  Generally, results from the diet testing have indicated that the sampled populations (through 2011) have been sensitive to the toxins and have responded similarly to control populations or CRW collected from previous seasons.  In some cases, lower susceptibility has been observed in field populations relative to the laboratory colonies (see data summaries in BPPD 2013a, b).  These results were likely due to natural variability, and the populations were not considered resistant. 

Table 3.  Summary of the mean susceptibility to Cry34/35 from diet bioassays of CRW populations collected 2004 to 2011[1] (Table generated from data provided by Dow AgroSciences and Pioneer).  

                                     Year
                               EC50 (ug/cm[2])
                               EC99 (ug/cm[2])
                               LC50 (ug/cm[2])
                               LC90 (ug/cm[2])
                                  2004-2005 
                                   0.8 - 1.1
                                       
                                   2.0 - 2.4
                                       
                                     2006 
                                   0.8 - 2.3
                                  21.0 - 56.6
                                   1.2 - 7.3
                                  9.2 - 78.1
                                     2007 
                                   1.0 - 3.7
                                 51.9 - 213.6
                                  3.0 - 11.5
                                  23.9 - 463
                                     2008 
                                   2.0 - 7.9
                                 83.8 - 596.6
                                  6.2 - 15.4
                                 58.4 - 613.6
                                     2009 
                               Field: 1.9 - 5.7
                                   Lab: 3.6
                                       
                              Field: 45.9 - 93.3
                                   Lab: 26.1
                                       
                                     2010 
                               Field: 2.4 - 7.6
                                   Lab: 2.9
                                       
                               Field: 6.7 - 33.7
                                   Lab: 14.9
                                       
                                   2011[2] 
                               Field: 2.6 - 5.3
                               Lab (Set A):  4.5
                               Lab (Set B):  2.4
                                       
                              Field: 11.7 - 39.9
                              Lab (Set A):  25.9
                              Lab (Set B):  12.4
                                       
[1] Results are for field populations collected by ABSTC; does not include results from CRW sampled from fields with unexpected damage or other populations.
[2] Testing in 2011 was conducted with new toxin lots (activity was similar to the previous lots).

Table 4.  Cumulative Results of mCry3A Susceptibility Testing for WCRW:  2006 - 2011[1] (Table created from data submitted by Syngenta).
 
                                     Year
                             LC50 range (ng/cm[2])
                             LC90 range (ng/cm[2])
                                 MRID Citation
                                     2006
                                357.9 - 1923.6
                               3788.7 - 28,766.6
                                   473401-01
                                     2007
                                209.8 - 1172.3
                               2470.5 - 52,515.0
                                   480470-02
                                     2008
                                191.9 - 4822.6
                           1430.0 - 37,820,880.0[2]
                                   480470-02
                                     Year
                             EC50 range (ng/cm[2])
                             LC50 range (ng/cm[2])
                                 MRID Citation
                                     2009
                            Field: 243.63 - 1137.07
                           Lab (French Ag.): 446.53
                           Lab (Crop Char.): 447.89
                           Field: 1297.98 - 15305.88
                          Lab (French Ag.): 14036.07
                           Lab (Crop Char.): 1084.21
                                   no MRID#
                                     Year
                            EC50 range (ug/cm[2])
                            LC50 range (ug/cm[2])
                                 MRID Citation
                                    20103 
                              Field: 0.19 - 1.53
                                   Lab: 0.42
                              Field: 1.49 - 13.96
                                   Lab: 3.82
                                   485913-01
                                     2011 
                             Field: 0.24  -  3.12
                                   Lab: 0.59
                             Field: 0.53  -  28.83
                                   Lab: 2.06
                                   489634-02
[1] Results are for field populations collected by ABSTC; does not include results from CRW sampled from fields with unexpected damage or other populations.
[2] One population in 2008 exhibited high tolerance to mCry3A.  The other sampled populations had susceptibility ranges of 191.9 - 1518.4 (LC50) and 1430.0 - 225,779.9 (LC90).
[3] Units of measure were changed from ng/cm[2] to ug/cm[2] starting with the 2010 testing.
                                       
Associated Uncertainties - Diet bioassays

As discussed in BPPD's reviews of monitoring data for Cry3Bb1 (BPPD 2011, 2012), Cry34/35 (BPPD 2013a), and mCry3A (BPPD 2013b), CRW are difficult to test using diet-based approaches.  Responses to toxin-incorporated diets are highly variable and a functional "diagnostic" concentration (capable of detecting resistant individuals) has not been developed for any of the registered toxins.  

To analyze diet bioassays, comparisons are typically made with susceptible control colonies and/or the historical range of susceptibility data for the toxins.  These comparisons can be complicated by variable responses from both field-collected and laboratory colonies (in some years, responses have varied by several orders of magnitude among field populations).  Much of this response can be attributed to natural variability in CRW and/or laboratory methodology, though it is unclear whether natural variation can be distinguished from a population that may be in the process of developing resistance.  Furthermore, data by Meihls (2010) suggest that LC50/EC50s of Cry3B resistant colonies may not always be distinguishable from those of the control populations, an indicator that present diet bioassay methodologies are not sensitive enough. Historical comparisons can also be limited if the testing methodology is modified or toxin lots are changed or sourced differently.  In addition, given the dietary differences between artificial diet and actual corn plants, a population showing relatively lower susceptibility in a diet bioassay may not necessarily be more fit or capable of surviving on Bt corn in the field.

The methodological issues with diet testing appear to be such that current resistance detection criteria may be unworkable in practice.  A number of field-collected populations have shown LC50s exceeding the upper limit of the 95% confidence interval of the control population (see data reviewed in BPPD 2011, 2012, 2013a, b, c), although none of the populations have been deemed resistant by registrants.  Presumably this is due to the difficulty of interpreting and obtaining accurate LC50 values; CRW are less sensitive to Bt than other insects and large amounts of toxin (or extrapolations from lower concentrations) are needed to obtain a 50% mortality. At sublethal concentrations, the insect ceases to eat and eventually dies from lack of food, which can lead to false positives when interpreting the diet bioassay results using LC measures.  The more conservative EC50 tests (requiring less toxin) have also shown some populations with reduced susceptibility (compared to control populations or the 95% CI of the historical range) and similarly have not been considered as evidence of resistance.

In addition to EC/LC measures, average body weight and head capsule width data of field populations raised on Bt diet has been compared to those of the susceptible control colonies raised on non-Bt diet. If there was a statistically significant difference exceeding the historical data for the controls, the result was interpreted as a case of suspected resistance. BPPD notes there is no evidence demonstrating that resistant corn rootworm need to have greater body mass or head capsule widths for development than their susceptible counterparts. Hence, this significance criterion is likely too stringent and would potentially result in false negatives. It seems more reasonable to expect that there will be no difference in body size between resistant (fed on Bt diet) and susceptible corn rootworm larvae (fed on non-Bt diet).

Unlike resistance monitoring for lepidopteran insects (see BPPD 2010e), discriminating concentrations are not used for CRW to distinguish potentially resistant and susceptible populations.  However, it is unknown whether this approach would be realistic given that the toxins are non-high dose and CRW may be less susceptible to Bt toxins due to selection pressures since commercialization.

Given the challenges identified with a diet bioassay-based paradigm, BPPD is not confident that a timely determination can be made to confirm resistance that would allow for a successful implementation of a remedial action plan. Since early resistance detection in corn rootworm is unlikely with the diet bioassay approach, growers and extension entomologists will likely recognize resistance in the field before currently established regulatory triggers will be met. To illustrate, it is unclear whether the Iowa and Illinois populations tested by Gassmann et al. (2011) could have been "confirmed" as resistant using the existing regulatory approaches.  In light of these concerns, BPPD has viewed "on-plant" assays as a potential solution to the need for a diagnostic CRW screening tool (see discussion in the next section).
  
b. On-plant Bioassays

Because of the inherent uncertainties (high variability, less sensitivity) with the present diet bioassay methodology for CRW, BPPD has recommended the use of a diagnostic on-plant assay as the primary resistance detection tool (such as those developed by Gassmann et al. 2011 and Nowatzki et al. 2008).  

Gassman et al. (2011) On-Plant Assay

Gassmann's assay (Gassmann et al. 2011) was specifically recommended for Cry3Bb1 follow-up investigations because it was developed with Cry3Bb1 corn as a diagnostic tool for resistance detection and would allow for comparisons of industry's results to the work of independent public research scientists (BPPD 2012). This particular on-plant assay utilizes Bt corn plants grown under greenhouse conditions to assess larval survival and represents a more natural feeding environment as opposed to artificial diets incorporated with purified toxin. This section describes the procedures used by Gassmann et al. (2011) to investigate Cry3Bb1-resistant populations, which could be adapted to other toxins.

Newly hatched neonates were distributed equally into the soil of containers planted with either one Bt or one non-Bt expressing corn plant and moved into an incubator (Temperature 25°C, 65% RH, 16/8 L/D), with each plant/container representing a replicate. After 17 days, larvae were extracted from the soil with a Berlese funnel. This duration allowed for development to last instar for the fastest developing larvae. The root masses were kept in the Berlese funnel for no less than four days during which time CRW larvae were continuously collected into vials containing 10 mL of 85% ethanol. The average sample size was 12.7 to 12.8 bioassay cups (number of replicates per treatment).

The proportion of survivors after 17 days was calculated for each cup by dividing the number of survivors by the number of neonates originally added to the soil of each cup. The mean proportion of survivors of each CRW population per treatment was calculated and analyzed with a two-way, mixed model analysis of variance. The fixed factors were they type of field (problem field or control), type of corn (Bt or non-Bt expressing) and interactions. Random factors were the "corn rootworm population" and "interaction between type of corn and population." Data were transformed to ensure normality and homogeneity of variance. Corrected mortality in the Bt cups was adjusted (using Abbott's correction method) by accounting for the average mortality observed in the non-Bt cups. Corrected survival was used to test for correlations with Bt and the number of years the population has been exposed to the Bt toxin.  Results from the study showed that populations collected from Cry3Bb1 "problem fields" (with unexpected damage) had significantly greater survival than populations from control (non-damaged) fields.  The researchers concluded that the data from problem field populations was evidence of "evolving resistance" to CryBb1 corn (Gassmann et al. 2011).

Associated Uncertainties - Gassman On-Plant Assay

At the 2013 USDA North Central Coordinating Committee (NCCC-46) meeting (New Orleans, LA) for corn rootworm research, some corn entomologists and industry scientists expressed concern with the potential standardization of an assay for all CRW toxins. There was discussion that more research is needed to evaluate which of the currently available on-plant assays (Gassmann et al. 2011 or Nowatzki et al. 2008) would be more appropriate for each of the registered Bt toxins. Both groups emphasized that the sublethal seedling assay approach (described in the next section) worked well for Cry34/35 because Pioneer had been optimizing the assay conditions and methods for several years. At the same meeting, some scientists from academia and industry contended that it was not entirely clear that the diet bioassay approach was obsolete and that more research is needed to investigate whether the assay could be further optimized for CRW resistance monitoring to distinguish resistant and susceptible individuals.

Dr. Gassmann proposed that a resistance ratio (proportion of corrected survival of field population / proportion of corrected survival of control population) could be used to determine whether a population is resistant. However, the ratio to determine resistance would need to be lower (i.e., > 4.0) for CRW than for other pests that are exposed to high dose toxins (resistance ratios > 10 have generally been associated with resistance  -  see Tabashnik 1994). This is because susceptible larvae in the field have likely been selected since the commercialization of Bt corn and may no longer be truly susceptible, which would effectively lower the resistance ratio by increasing the denominator (survival of the control population). It is also uncertain how representative current laboratory colonies (reared in absence of Bt) are to truly susceptible wild types that have been exposed to the widely-adopted Bt corn PIPs in the field.

Sublethal Seedling Assay - Nowatzki et al. (2008) 

A second on-plant testing approach, the Sublethal Seedling Assay (SSA), has been developed by Pioneer scientists for the Cry34/35 toxin and is described in Nowatzki et al. (2008).  The testing methodology for the SSA involves exposing neonate CRW to seedling mats of Bt or non-Bt corn.  As implied by the name, the SSA does not measure mortality; rather, the test is designed to assess potential developmental effects of Bt exposure including growth (instar) and body size (larval body area).  Potential resistance can be detected by statistical comparisons of larvae exposed to Bt and non-Bt corn treatments (resistant larvae are assumed to develop as well on Bt corn as non-Bt corn).  

Pioneer has been using the SSA as part of the Cry34/35 monitoring program since 2009 in conjunction with diet bioassays.  As the company has gained experience with the assay, the methodology and data analysis have been modified (see discussion in BPPD 2013a).  The general procedure for the SSA (modified from Nowatzki at al. 2008) is described in the 2012 resistance monitoring submission to EPA (see MRID# 490085-01, as discussed in BPPD 2013a) and is briefly summarized here. Each seedling mat (150 seed kernels per mat) is infested with approximately 500 CRW eggs on the soil surface (maintained in an environmental chamber at 25 °C, 65% RH, 14:10 light/dark).  A separate hatch test (using 100 egg samples in Petri dishes held in the same environmental chambers) is conducted to determine the starting point of the test.  After the initiation of egg hatch, larvae are exposed to the root mats for 17 days (~ 154 degree days) to allow for growth and development.  Larvae are then collected from the root mats using Burlese funnels positioned over containers with 70% ethanol.  Samples are processed by pouring the ethanol containers into collection pans with numbered grids.  The total number of larvae is counted in each pan and a subset of 30 is randomly collected by using a random number to pick an individual grid square (repeated until 30 can be collected).  These larvae are then measured (head capsule width) to determine the developmental instar (1[st], 2[nd], or 3[rd]).  A second processing procedure is also used to assess body size.  Samples are poured into Petri dishes, which are scanned and analyzed by imaging software to calculate the total body area (mm[2]) of each larva.

Data from the SSA testing are analyzed using several different statistical methods.  For the instar data, Pioneer analyzes the distribution of larvae among all three instars.  A mean instar is calculated for each treatment group as follows:  

mean instar=1*#1st instars+2*#2nd instars+3*#3rd instarstotal # larvae

Bt and non-Bt treatments are compared to generate a means ratio (mean instar of Bt / mean instar of non-Bt).  If larval development on Cry34/35 equals or exceeds that on non-Bt, the ratio will be >= 1.0.  The instar data are statistically analyzed by generating a Kolmogorov-Smirnov statistic (KS D) that accounts for the magnitude of difference between treatments (ranging from -1 to 1, with 0 indicating no differences between distributions).  Instar distribution is also analyzed using the Fisher's exact test.  [In earlier testing with the SSA, Pioneer calculated an odds ratio for each treatment based on the proportion of larvae reaching the third instar.  However, shortcomings with this approach became apparent with the assays conducted during 2010.  The third instar odds ratio failed to account for differences between first and second instar development and was hampered by treatments with low larval counts.] 

Body size data are analyzed in a comparable manner.  Mean body area measurements are compared between treatments to create a ratio (body area of Bt exposed CRW / body area of non-Bt exposed CRW).  Similar to the instar data, a KS D statistic is employed to assess differences between the treatment groups.

Pioneer's decision process to determine suspected resistance based on the SSA (i.e., triggering follow-up investigation) includes two criteria.  Further action will be taken if a population:
  
 Shows similar response when exposed to both 59122 (Cry34/35) and isoline at a significance level of p>0.05 for both methods (instar and image data); 
 Shows mean ratios (both instar and size) outside the 95% confidence interval established from data collected (in) previous years.  Additional SSA data (i.e., from the 2012 sampling/2013 testing) will be needed for establish the 95% confidence interval for this criterion.

Pioneer has conducted SSA testing on WCRW collected in the 2009, 2010, and 2011 seasons (data for populations collected in 2012 are due to be reported in Fall 2013).  These assays were performed on the same field populations collected by ABSTC for the diet bioassay assessments (see discussion in Section II above).  In addition to the field populations, Pioneer has included a Cry34/35-susceptible colony (obtained from USDA-ARS, Brookings, SD) and a Cry34/35-resistant colony (beginning in 2010) as reference populations.  Results from the SSA work is summarized in Table 5 below.

Table 5.  Sublethal Seedling Assay Results for Western Corn Rootworm, 2009 - 2011 (Table created from data submitted by Pioneer; MRID#s 486475-01, 486951-01, 490085-01)
                                     Year
                                 # Field Pop.
                        Instar (mean ratio Bt : non-Bt)
                      Body Area (mean ratio Bt : non-Bt)


                                     Field
                            Susceptible Control[1]
                              Selected Colony[2]
                                     Field
                            Susceptible Control[1]
                              Selected Colony[2]
                                     2009
                                      15
                                 0.766 - 0.996
                                     0.781
                                      ---
                                 0.393 - 0.894
                                     0.508
                                      ---
                                     2010
                                       7
                                 0.734 - 0.960
                                     0.787
                                     0.915
                                 0.387 - 0.511
                                     0.512
                                     0.692
                                     2011
                                     11[3]
                                 0.687 - 0.883
                                   1) 0.727
                                   2) 0.830
                                     0.986
                                 0.214 - 0.421
                                   1) 0.198
                                   2) 0.274
                                   1) 0.624
                                   2) 0.633
1 Cry34/35-susceptibile control colony obtained from USDA-ARS (Brookings, SD).
[2] Colony (York-RR) has been selected on 59122 (Cry34/35) corn.  
[3] Includes three populations collected from fields with unexpected CRW damage.  The results from these populations were within the ranges of the field-collected populations.

In addition to WCRW, Pioneer was able to test one NCRW population in 2009 (collected from Delaware, IA); no successful NCRW collections were made in 2010 and none of the five populations sampled in 2011 produced sufficient eggs.  The mean instar ratio for the 2009 population was 0.850 (KS D = 0.389; p < 0.0001) while the mean body area ratio was 0.598 (KS D = 0.406; p < 0.0001).  A susceptible NCRW control population was not tested.

Associated Uncertainties - SSA

Pioneer's 2009 - 2011 reports demonstrated that the company has made significant progress towards developing the SSA as a diagnostic screening tool for Cry34/35 resistance.  As with any bioassay, BPPD believes a major question will be how the results are statistically analyzed and interpreted in the context of a resistance determination. 

The approach to detecting resistance with the SSA involves positive results (i.e., statistically similar responses of a CRW population to Bt and non-Bt corn) with two measures (instar development and body size).  As described above, under Pioneer's "decision process" for the SSA populations that respond (statistically) similarly on Bt and non-Bt plants to both measures are to be classified as "suspected" resistant.  The company stated that follow-up action will be taken with populations meeting the criteria, though it is unclear what additional testing (if any) would be employed and what time frames would be needed.  BPPD concluded that both the bioassay metrics and decision interpretation process for the SSA are reasonable (see discussion in BPPD 2013a), but recommended a conservative approach, with follow-up analysis on any population that could reasonably be considered resistant to the toxin.  Given the design of the SSA, BPPD recommended that the decision process be used to support a conclusion of "resistance" (instead of the more ambiguous "suspected resistance"), with any population meeting the threshold subject to the appropriate remedial action plan. (A more detailed discussion of resistance definitions and criteria is found in Section IV.)

Assay sensitivity is another consideration for the SSA. As part of the 2010 - 2011 testing, Pioneer included a population that was laboratory-selected on Cry34/35 (termed "York-RR" in the 2011 report).  BPPD (2013a) presumed that this population has resistance to the toxin and was derived from the "York" colony described in Lefko et al. (2008).  The study reports, however, did not confirm this assumption or include any quantification of the population's susceptibility (or resistance) to Cry34/35.  In the 2010 testing, the York population's mean instar and body area measures were statistically lower on Bt plants than non-Bt.  For 2011, the York-RR colony had statistically similar mean instar values (Bt vs. non-Bt) but mean body area was significantly less on Bt compared to non-Bt.  In both of these years, the York population appears not to have met the criteria necessary to be considered "suspected resistant" to Cry34/35 according to the decision process proposed by Pioneer.  To the contrary, the company concluded that the 2011 results "clearly identified the York-RR laboratory selected colony as resistant against 59122."  But the study report did not address the mean body area size results (which are considered equally in the determination of resistance) that appeared to show no resistance (though the Bt-exposed larvae were larger than the field-collected populations).  

While the specific level of resistance of York-RR to Cry34/35 is unclear, these results raise questions regarding the sensitivity of the SSA to detect resistant field populations.  If resistant populations are not identified (i.e., false negatives), the assay will be limited in its ability to screen for tolerance to Cry34/35.  This may be a particular concern for populations sampled from Cry34/35 fields with unexpected CRW damage (three such populations were tested in 2011).  As discussed above, a more conservative (or flexible) approach to interpreting the data may be needed.  Further refinement of the assay techniques (using the selected colony) may also help improve sensitivity.

IV. RESISTANCE CONFIRMATION  

a. Current Procedures and Definitions of Resistance

There are three regulatory criteria incorporated in the terms and conditions of each Bt PIP registration that, when met, will determine resistance in corn rootworm. These triggers are: "unexpected damage" threshold, "suspected resistance," and "confirmed resistance." The process to determine resistance establishes sequential steps that aim to rule out other causes for observed damage and to decrease the likelihood of false positives (reporting resistance when it has not occurred). There is concern, however, that actionable thresholds may not be met until resistance is widespread and effective mitigation (reducing the resistance allele frequency) is impractical.  In this case, managing resistance through population suppression may be the only alternative. 

When a company receives and investigates a report of field damage, the first steps are to rule out other causes of observed damage and test for the presence and expression levels of the Bt toxin. Provided the toxin is present at the expected protein expression levels in the lodged plants and other causes are not implicated, root damage is assessed using the Iowa State Nodal Injury Scale (NIS; Oleson et al. 2005).  The "unexpected damage" thresholds (average damage for the field) are 1.0 for single toxin products and 0.5 for pyramided toxins (with > 50% plants damaged); if these levels are exceeded, resistance is "suspected", and the registrant must then collect adult corn rootworms in the damaged fields. Reports of unexpected damage are often received when adult corn rootworm are no longer present in the corn fields, which makes adult insect collections impossible. In such cases, adult collections would have to be made the following year in the same vicinity, though alternate CRW control measures (e.g., insecticide use) may lessen CRW presence and damage on Bt corn plants. After adults are collected, mated, and eggs harvested, diet bioassays (or more recently on-plant assays) are conducted with offspring from field collections to evaluate whether confirmed resistance exists. If resistance is suspected, registrants are generally required to work with affected growers to implement mitigatory measures for CRW control, such as adulticide treatment, crop rotation the following year, or use of soil or seed insecticides the following year. Some of these measures are intended to eliminate potentially resistant insects from contributing to the following year's CRW population; others are management practices intended to suppress population densities and to reduce root node injury damage the following year (see Section VI for more information).

The current regulatory language has varied among registrations with respect to the definition of "confirmed resistance." This is because each company proposed their own criteria of "resistance" and worked together with the EPA to have those established in their terms and conditions of the PIP registrations.  In some cases (e.g., Cry3Bb1), the definition of resistance has been modified since initial registration.  Confirmed resistance is defined as follows:



Cry34/35 (Pioneer, Dow) and mCry3A (Syngenta) 

Single toxin Cry34/35 or mCry3A products (paraphrased from the terms of registration  -  see BPPD 2010a)

Resistance is confirmed if all of the following criteria are met by progeny from the CRW sampled from the area of suspected resistance:

      (a) The proportion of larvae that can feed and survive on (Cry34/35 or mCry3A) roots from neonate to adult is significantly higher than the baseline proportion; 
      
      (b) The LC50 of the test population exceeds the upper limit of the 95% confidence interval for the LC50 of a standard unselected population and/or survival in the diagnostic assay is significantly greater than that of a standard unselected population, as established by the baseline monitoring program;
      
      (c) The ability to survive is heritable; 
      
      (d) (Cry34/35 or mCry3A) plant assays determine that damage caused by surviving insects would exceed economic thresholds;
      
      (e) The identified frequency of field resistance could lead to widespread product failure if subsequent collections in the affected field area(s) demonstrated similar bioassay results.

Pyramided products with Cry34/35 and mCry3A (paraphrased from the registration terms of EPA Reg. No. 29964-16 and 67979-20)

In 2012, EPA registered several pyramided products containing Cry34/35 and mCry3A.  As part of the registration approvals, the terms for CRW resistance monitoring were modified from previous definitions.  For these products, resistance is confirmed if all of the following criteria are met:

 Injury potential of the field-collected rootworm population feeding on plants containing Cry34/Cry35Abl and mCry3A remains at a level likely to produce repeated product failure in field conditions;
      
 Subsequent populations collected from the area and assayed show that the results are repeatable;
      
 The change in injury potential has been documented as a heritable characteristic of the target pest population;
      
 Greenhouse node-injury evaluation confirms product failure; and
      
 Continued monitoring of the area suggests that the change is spreading.

Cry3Bb1 (Monsanto) 

Original definition (paraphrased from the Cry3Bb1 BRAD -- BPPD 2010b)

Collected CRW will be reared to adults and mated amongst themselves or single-pair mated with individuals from a susceptible lab colony if numbers are low. The resulting progeny will be exposed to the diagnostic concentration bioassays to determine heritability of survival on a Cry3Bb1-containing diet. If heritability is confirmed, survivors will be placed on MON 88017 corn plants to assess whether level of resistance is enough to cause severe root damage. Whether an increase in susceptibility has occurred will be assessed with a discriminating concentration bioassay when such a dose has been established. Until then, either of the following criteria below will serve to confirm resistance:

      (a) "The LC50 of the standard bioassay exceeds the 95% confidence interval of the mean historical LC50 for susceptible pests according to the baseline measurement"; or

      (b) "Over 50% of Cry3Bb1-expressing plants have >=1 root nodes destroyed by suspected resistant populations under controlled lab conditions."

Revised definition (paraphrased from the registration terms of MON 88017, EPA Reg. No. 524-552)

In response to concerns raised by BPPD regarding Cry3Bb1 resistance monitoring (see BPPD 2011 and 2012), the process to determine resistance was revised in 2013.  Monsanto amended all single toxin Cry3Bb1 products (EPA Registration Numbers 524-552, 524-576, and 525-606) to reflect the new approach.

Monsanto must conduct investigations of all CRW populations collected as part of the performance inquiry process (i.e., investigations of unexpected damage). A CRW population will be considered resistant to Cry3Bb1 if the following criteria are met and additional collections and testing are not deemed to be necessary (see below): 

 An initial performance inquiry investigation results in a finding of Unexpected Damage; and
      
 Where green tissues are available and plants are unusually stressed, Bt protein levels in affected plants are found to be within the documented range for that hybrid (if data are available); and
      
 Single, on-plant bioassays of insect collections from the affected fields show the following:
      
 A statistically significant difference in measures of either lethality/mortality or sublethal effects (growth/development) between the field population and the control population on Bt corn and
            
 A lack of a statistically significant difference in the same measures of the field population raised on Bt corn and non-Bt corn plants.

To minimize the potential for incorrectly reaching a conclusion of resistance, another year of CRW adult collections and additional testing would be needed to determine resistance if:

 The results of the single, on-plant bioassays are inconclusive (e.g., the results of the statistical analysis are unclear because of low sample sizes) or
      
 Another reasonable explanation for the unexpected damage exists (e.g., high pest pressure and/or high plant stress).
      
In these cases, Monsanto and EPA will discuss and align on next steps before any resistance conclusion is reached.
	
If CRW collections are not possible in the current year or subsequent years due to successful management practices, then no further investigation is needed. The population would be considered "mitigated" meaning, in this case, that the population is suppressed or extirpated in this location. However, EPA recommends that Monsanto continue to be vigilant in areas where CRW populations were successfully mitigated.  

Associated Uncertainties  -  Current Procedures and Definitions of Resistance

The use of root node injury scores (NIS) has been a critical screening process in the current monitoring program for determining potential resistance.  However, root node injury from larval corn rootworm feeding can be enhanced by environmental conditions as well as corn rootworm larval density present in the soil. The challenge for effective resistance monitoring is therefore to determine a threshold of root node injury that 1) reflects decreased susceptibility indicative of putative resistance in corn rootworm but 2) is not too conservative as to miss a resistance event (false negatives) or too sensitive as to initiate unnecessary "suspected resistance" investigations (false positives). False negatives are counterproductive to early resistance detection and possibly limit remediation measures, while too many false positives can place unnecessary burdens on industry and EPA. The use of crop damage ratings are further complicated by the fact that none of the currently-registered Bt toxins for CRW is considered high dose.  Consequently, a portion of CRW larvae can be expected to survive exposure to Bt corn even if they are not resistant to the toxin (SAP 2002).

At present, there are two sets of thresholds for "unexpected damage" fields that trigger further investigation by the registrants (i.e., "suspected resistance").  The thresholds, based on average root damage for the field, are >1.0 for Bt corn expressing single toxins and 0.5 for Bt corn products expressing two or more (pyramided) Bt toxins.  These standards have been subsequently adopted by Bt corn registrants, though for some pyramided products the 0.5 damage level must be observed on at least 50% of sampled plants.

While the use of NIS scores is helpful for identifying less susceptible populations for further testing, BPPD cautions against using root node injury as a criterion for determining actual resistance. Root damage can be exacerbated by environmental factors, and excessive node injury scores observed on roots of corn plants are not always reflective of the density of corn rootworm and/or resistance in a population. Gray (2012) reported, for example, that root injuries to several corn rootworm Bt hybrid traits have in some cases been observed shortly after commercializaiton of the traits began. These observations suggest that root damage can not always be linked to resistance development. 
   
The timing of the resistance detection process presents additional uncertainty.  Under ideal circumstances CRW are collected from damaged fields during the same season as the report and reared for testing.  But obligate and extended diapause in CRW delay the bioassays into the next season and results may not be available until a year after the initial collections.  If adults could not be sampled due to late season timing or mitigation (beetle bombing), collections are attempted the following year, further delaying a resistance determination.  Additionally, great variability in response to diet bioassays and lack of a functional diagnostic dose hinder the ability to make a timely conclusion of "confirmed resistance." 

Overall, resistance monitoring for CRW is reactive instead of proactive.  Resistance is more likely to be detected from damaged fields than from the randomly-sampled populations that are tested for shifts in susceptibility relative to historical data (prior to field failure).  Because of this, the ability to successfully mitigate a resistant population (by containment) may be limited.  Best management practices (such as crop rotation, use of different mode of action, etc.) likely need to be implemented for the following year irrespective of whether resistance testing has been completed. Recent registrations have addressed this in part by including requirements to implement BMPs if unexpected damage thresholds are met (i.e., "suspected resistance").

As discussed above, the resistance detection process for CRW has involved a series of steps including "suspected" and "confirmed" resistance.  Populations suspected of resistance must still be tested to determine whether resistance has actually evolved.  To avoid confusion over the use of the word resistance, the term "suspected resistance" was dropped from Monsanto's Cry3Bb1 registrations in 2013 (as part of enhancements to the CRW resistance monitoring protocol).  Instead, CRW populations meeting unexpected damage thresholds are termed "populations of interest" for further testing; only such populations that meet the decision criteria in subsequent bioassays are deemed "resistant."

b. Proposed New Definitions of Resistance

Given the scientific uncertainties associated with artificial diet bioassays (see section III), BPPD has proposed the use of the single on-plant and/or sublethal seedling assays as the primary diagnostic tool to determine if a CRW population is resistant to Bt (BPPD 2012).  Both on-plant assays have greater ability to separate susceptible and resistant populations than the diet bioassay methods used to date. 

Interpreting the results of on-plant assays is critical for any conclusions regarding resistance.  On-plant assays can address a number of metrics (e.g., survival, instar development, body size and weight) that can be used to evaluate Bt susceptibility. Comparisons can be made 1) between a sampled population's response on Bt corn and non-Bt corn or 2) between the sampled population and a (susceptible) laboratory population.  BPPD believes that assessing two or more of these measures in an on-plant assay will add robustness to the resistance determination and reduce the potential for false positives, while providing a more proactive resistance monitoring tool.  

Bt corn registrants and BPPD have identified several potential new options for defining resistance, based on the use of on-plant assays.  Separate approaches have been proposed for the Sublethal Seedling Assay (Nowatzki et al. 2008) and the single on-plant assay (e.g., Gassmann et al. 2011 or similar methodology).  The definitions utilize multiple comparisons/susceptibility measures, each of which need to be fulfilled before resistance can be demonstrated. Populations obtained through routine annual resistance monitoring efforts as well as from Bt corn fields reported to have had unexpected damage (with average root node injuries greater than >1.0 NIS for single toxin Bt corn and >0.5 NIS for pyramided Bt corn) should be subjected to this analysis and approach.

Sublethal Seedling Assay Option 1 (developed by Pioneer for Cry34/35 monitoring  -  see Section III of this document):

   Using the sublethal seedling assay (Nowatzki et al. 2008), a corn rootworm population is determined to be resistant when:
 The sampled population shows similar response when exposed to both Bt corn and isoline corn at a significance level of p>0.05 for both instar and body size data; and
 The sampled population shows mean ratios (both instar and size) outside the 95% confidence interval established from data collected in previous years.
   
Sublethal Seedling Assay Option 2:

   Using the sublethal seedling assay (Nowatzki et al. 2008), a corn rootworm population is determined to be resistant when:
 The distributions (ratios) for instar development of the susceptible reference strain on Bt corn and the sampled field population reared on Bt corn demonstrate a statistically significant difference (P-value <0.05); and 
 The average weight (or body area) of the field population on Bt corn is equal to (no statistically significant difference) or greater than the average weight (or body area) of the susceptible colony on isoline corn.

Single On-Plant Assay Option 1 (employed by Monsanto for Cry3Bb1 monitoring starting with the 2013 season  -  see Part a in this section): 

   A corn rootworm population is said to be resistant if single, on-plant bioassays of insect collections from the affected fields show the following:
 A statistically significant difference in measures of either lethality/mortality or sublethal effects (growth/development) between the field population and the control population on Bt corn; and
 A lack of a statistically significant difference in the same measures of the field population raised on Bt corn and non-Bt corn plants.

Single On-Plant Assay Option 2: 

   Using a single on-plant assay (e.g., Gassmann et al. 2011 or similar methods), a corn rootworm population is said to be resistant when:
 The percent survival observed of the field population on Bt corn exceeds the percent survival of the susceptible population on Bt corn (statistically significant; P-value <0.05); and 
 The average weight of the survivors from the field population on Bt corn is equal to (no statistically significant difference) or greater than the average weight of the susceptible colony on isoline corn.

Associated Uncertainties  -  Proposed New Definitions of Resistance

The nature of resistance for corn rootworm and Bt toxins is not completely understood, though it is likely to affect how bioassay results are interpreted for resistance determinations.  A series of laboratory/greenhouse selection experiments with Bt toxins have shown that CRW resistance is unlikely to follow a single gene, recessive model known for other target pests of Bt crops with high dose expression (e.g., tobacco budworm, Gould et al. 1995).  Instead, research has suggested that CRW resistance may be non-recessive (Meihls et al. 2008) or could involve multiple genes conferring varying levels of tolerance (Lefko et al. 2008).  On-plant assays essentially function as a diagnostic screen that will identify populations with genetic resistance (the ability of offspring to survive on Bt plants), but may not detect incomplete resistance or other forms of resistance such as avoidance of the toxin by selective feeding on roots with differential toxin expression. 

Sublethal Seedling Assay Option 1 above is the approach currently employed by Pioneer to monitor Cry34/35.  Section III of this document contains a discussion of the assay and uncertainties with regard to resistance detection.  This approach incorporates two measures of growth (instar development and body area) and includes comparisons with historic data (which would need to be developed for toxins other than Cry34/35). A second proposed SSA method (Option 2) involves two comparisons with susceptible laboratory colonies.  The first is a comparison of response to Bt corn while the second compares the field colony response to Bt with the susceptible colony on isoline corn.  BPPD notes, however, that the second criterion is essentially a comparison of independent events (lab colony response on isoline and field colony response on Bt) that may not provide a reliable indication of potential resistance. 

Single On-Plant Assay Option 1 was integrated into the resistance monitoring requirements for Monsanto's Cry3Bb1 registrations in 2013.  Two levels of comparison are required by the procedure to confirm resistance:  the field colony vs. the lab colony on Bt corn and the field colony on Bt vs. non-Bt corn.  The second proposed single on-plant technique (Option 2) incorporates both lethal (survival) and sublethal (larval weight) effects from Bt corn as part of the resistance assessment.  This approach may be problematic, however, because the second criterion is not independent from the first. The larval weight measure would be obtained from survivors of the Cry3Bb1 treatment in the first criterion.

All of the proposed procedures are likely to detect resistance only if it is complete (i.e., CRW develop and survive on Bt corn at the same levels as isoline corn).  Comparisons of field populations on Bt and non-Bt corn require statistical equivalence to make a positive resistance determination.  This statistical approach would exclude cases of incomplete resistance, in which a field population is more fit than a control colony on Bt plants but not as fit as a population with complete resistance.

BPPD also notes that the proposed definitions could be further complicated if there is cost to resistance in the field population(s). Potentially, the survival of a resistant corn rootworm population could be significantly (statistically) lower on non-Bt corn than the survival of a susceptible population. Though cost to resistance with respect to decreased survival on non-Bt corn has not been demonstrated yet for corn rootworm, it should be considered in the resistance detection methodology.  A cost to resistance could be beneficial for insect resistance management mitigation in that the planting of isoline (non-Bt) corn may decrease resistance allele frequency in the population.

A number of the potential resistance definitions utilize body weight data as one of two criteria for resistance. However, body weight data for resistant field populations identified from on-plant assays has until recently not been characterized. BPPD has received preliminary on-plant assay data from Cry3Bb1 problem fields indicating that body mass of such populations can be equal to or greater than the body mass of the susceptible lab colony when tested on Bt corn (MRID 489283-01, see BPPD 2013c). But it is not entirely clear from this comparison how the weight data from Bt survivors would indicate when a population has evolved resistance. 

Regardless of the resistance definition that is adopted, appropriate statistical tests for each criterion in the definition will need to be established.  Statistical tests may vary depending on the variables measured in the assay and sample size will affect statistical power.  For the Sublethal Seedling Assay, Pioneer has employed a Kolmogrov-Smirnov (KS D) statistic to test for the magnitude of difference between treatments.  For other assays (e.g., Gassmann et al. 2011), analysis of variance (ANOVA) is frequently employed.  The level of significance in the statistical analysis should be set to reduce the likelihood of false positives, but not so low as to risk false negatives.  Typically, a significance level (p-value) of 0.05 has been utilized in resistance monitoring assays.  

The use of resistance ratios (i.e., the response of a putative resistant population relative to a susceptible control) may also be helpful to consider as a potential threshold for resistance, though several factors could affect their use.  A resistance ratio of 10 fold or greater has been used to define genetic resistance (Tabashnik 1994), but this standard may be difficult to obtain for CRW if proportion-based measures are assessed in the assay (e.g., percent mortality or larval recovery).  CRW are less susceptible to Bt than other insects (none of the toxins are "high dose"), and survival/recovery of 10% or more of susceptible larvae on Bt corn has been observed in some laboratory/greenhouse tests (e.g., Gassmann et al. 2011, Oswald et al. 2011).  Additionally, CRW populations in the field have experienced selection to Bt corn over the past decade and likely have obtained some degree of tolerance to the toxins (Cry3Bb1, mCry3A, and Cry34/35).  The combination of these factors could reduce resistance ratios below the 10 fold standard, even if 100% survival (i.e., complete resistance) is observed in the field population. 

Both the Sublethal Seedling Assay (Cry34/35) and single on-plant assay (Cry3Bb1; Gassmann et al. 2011) were developed for specific Bt toxins.  Adapting the assays to other toxins may present technical challenges in adapting certain Bt corn hybrids to the experimental design.  In particular, BPPD is aware of problems developing the SSA for mCry3A (see BPPD 2013b) and Cry3Bb1 (see BPPD 2013c).

c. Registrant Investigations of CRW Populations from Damaged Bt Fields

All Bt corn registrants (Monsanto, Pioneer, Dow, and Syngenta) have reported investigations of unexpected CRW damage to Bt corn.  Details of these investigations are described in the individual reviews for Cry3Bb1 (BPPD 2013c), Cry34/35 (BPPD 2013a), and mCry3A (BPPD 2013b) and are briefly summarized below.

Monsanto collected three populations from problem fields in 2011 for susceptibility testing.  Eleven additional populations were taken from fields that experienced high corn rootworm damage during 2010 but were not sampled at the time. Of the total eggs collected from the 14 populations, only 12 populations produced enough eggs to conduct artificial diet bioassays. Of the populations tested, four populations had mean EC50s that fell within or below the historical range (2007-2010) for diet bioassays and were comparable to the results obtained for the laboratory control treatments. One population (Schutte, CO) had a mean EC50 that exceeded the historical range.  EC50 values could not be estimated for the other seven populations.  Monsanto also assessed survival and body mass at the highest concentration tested (170.8 ug/cm[2]). Three populations (Humphrey, NE, Trent, SD, and Chatfield, MN) had lower average percent mortalities and higher average weight/survivor than the historical range.  For eight other populations, the average weight/survivor was above the historical control but the percent mortality was within the established range. Monsanto concluded that the results collectively indicated that eight populations (Maynard, IA, Shell Rock, IA, Onslow, IA, Schutte, CO, Chatfield, MN, Mankato, MN, Humphrey, NE and Trent, SD) of the 12 populations obtained from damaged fields had potentially reduced Cry3Bb1 susceptibility compared to the historic data and other population sampled in 2011.

A subset of the populations collected by Monsanto was also tested using a single on-plant assay. Six populations were assessed for mortality and developmental effects (head size and weight) on CryBb1 and non-Bt corn plants.  Two of these populations remained highly susceptible to Cry3Bb1.  The other four (Shell Rock, IA, Onslow, IA, Trent, SD, and Mankato, MN) had statistically similar larval survival on Bt and non-isoline corn, though percent survival was numerically lower on Bt for three of the populations. One of the populations (Trent, SD) also had statistically similar Bt vs. non-Bt responses for head width and weight.  Monsanto concluded that these results were generally in agreement with the results obtained from diet bioassays (MRID# 489283-01).

Dow and Pioneer tested one population from 2009 and four populations from 2011 that were collected from Herculex Xtra fields with unexpected CRW damage (defined by root damage rating in excess NIS 1.0).  Bioassay results for the 2009 population were within the range of other field collected CRW (from non-damaged fields).  The four populations sampled in 2011 generally had lower EC50 and LC50 susceptibility than the ABSTC field populations.  CRW were also sampled from nearby undamaged "comparison" fields for testing; the results from these fields were similar to the unexpected damage fields.  Pioneer further tested the unexpected damage samples using the Sublethal Seedling Assay and found that the populations' susceptibility (instar and body size rations) was comparable to the ABSTC-sampled populations.  The other 2011 population from an unexpected damage site (Knox County, NE; sampled by Dow) had the highest LC50 of all populations tested but was not further investigated using the SSA.  Dow reported that the field was rotated to soybean in 2012 (MRID#s 482797-01, 490266-02).

In its 2011 report, Syngenta indicated that 12 CRW populations were collected from fields with unexpected damage (locations of the damaged fields were not described).  Bioassays were in progress at the time of the report, and no data have yet been made available to EPA (MRID# 489634-02).  

V. NORTHERN CORN ROOTWORM MONITORING

As part of the 2010 Bt corn registration extensions (BPPD 2010d), EPA required registrants of CRW-targeted products to develop and implement a resistance monitoring program for northern corn rootworm (NCRW).   The specific registration term read as follows:  

      "[Registrant] must develop a proactive resistance monitoring program for northern corn rootworm (Diabrotica barberi) a program by the 2012 season, with annual reporting in 2013.  This program should include a proposal for annual sampling and testing of northern corn rootworm susceptibility to [toxin].  As part of this effort, [registrant] may need to investigate novel techniques for rearing and conducting bioassays with northern corn rootworm. A report on the [registrant's] progress towards this requirement must be submitted within one (1) year of date [September 29, 2010] of the amended registration." 

To address this requirement, ABSTC conducted a workshop in November 2011 to "assess the potential for a reliable NCR resistance monitoring program" and would submit a report from the meeting to EPA.  The ABSTC workshop included scientists from Monsanto, Syngenta, Pioneer, Dow, and DM Crop Research Group as well as two academic researchers (Dr. Lance Meinke, University of Nebraska, and Dr. Ken Ostlie, University of Minnesota).  Topics discussed at the meeting included sampling/beetle collection, rearing/egg production, and bioassays. 

ABSTC's report (letter to EPA dated August 12, 2012) indicated that "the workshop concluded that the current technical and logistical constraints make it impossible to reliably monitor northern corn rootworm for changes in susceptibility over time."   The minutes identified a number of specific challenges in working with NCRW relative to WCRW.  Maintaining and transporting NCRW colonies is difficult and egg production can be erratic.  NCRW are sensitive to crowding, pathogenic and environmental stresses that can negatively impact laboratory populations.  Though improved by the experiences of USDA-ARS with the Brookings (SD) colony, further optimization of NCRW rearing techniques will likely be needed.  In terms of bioassays, the workshop attendees noted that test methods developed for WCRW would need to be validated for NCRW.  Diet and on-plant assays that have been used for WCRW (e.g. SSA) could also be potentially used with NCRW.  A susceptible laboratory control is needed but it was unclear whether the Brookings NCRW colony would be sufficient for all of the testing needs.  Furthermore, there were concerns that extended diapause (observed in some NCRW populations) could hamper the timing for conducting the bioassays.

An action plan for 2012-2013 was developed in the workshop.  Three objectives were discussed:  1) consultation with USDA on the availability of the Brookings NCRW colony and the possible development of a non-diapausing line; 2) potential collection of NCRW in 2012 with the use of trap crops; and 3) proof of concept testing with each company's bioassay using the same NCRW laboratory colony.  ABSTC's report indicated that the group would inform EPA on the progress of the NCRW resistance monitoring effort.

BPPD recognizes that NCRW present significant logistical challenges for resistance monitoring, including difficulties with sampling, rearing, egg production, and testing with bioassays.  For this reason, BPPD concurs with ABSTC's conclusion that at the present time it is not possible to effectively monitor NCRW for changes in toxin susceptibility. 

BPPD notes that one NCRW population was successfully tested by the SSA (see section III in this document), an indication the assay may hold promise as a practical resistance detection tool.  The greater challenge with NCRW, however, appears to be in sampling and laboratory rearing.  ABSTC attempted to make multiple NCRW collections during 2009-2011, but of the total 24 locations identified for sampling only 10 resulted in successful beetle collections.  Of those, just one population from 2009 produced a suitable number of eggs for testing by SSA. 

VI. REMEDIAL ACTION

a. Remediation
 
The term `Remediation' is typically used to describe an action that reverses or stops an environmental damage.  Remediation in the context of IRM has been broadly interpreted and taken on at least two meanings, one referring to reversal of resistance and the other to suppression of resistance. For example, remediation is used to describe efforts aimed at decreasing the resistance allele frequency in a population(s) and containing resistance geographically. This could be achieved by rotating from Bt corn to a non-host crop that does not support the target pest (i.e., soybean for corn rootworm) and extirpating the population(s). If cost of resistance has been identified in the target pest, then planting non-Bt corn would be another remediation strategy that reduces resistance in the population. Remediation has also been used to describe management efforts that suppress insect numbers of a resistant population to a point where damage in the Bt field no longer leads to economic injury levels. Applications of soil insecticides to control corn rootworm larvae and/or spraying of adult corn rootworm beetles are examples of such suppression tactics. Such management practices do not actually reduce the overall resistance because resistant and susceptible individuals in the population are equally likely to die from the pesticide applications. If such a field is no longer managed, then insect densities will likely build up again and resistance could be evident as Bt crop failure. The BPPD IRM team contends that population suppression is not actually remediation (rather a management practice) unless the population is extirpated and the affected fields are managed as resistance `hot spots' (see part b below). 
 
The terms and conditions for all Bt corn registrations have a generic remedial action plan in place, which is to be superseded by a registrant's specific remedial action plan (submitted to the Agency within 30-90 days) if resistance is confirmed. The aim of the specific remedial action plan is to uniquely address each resistance situation (i.e., nature of target pest, PIP, extent of geographic area affected). The general aspects of the current remedial action plans (in response to "confirmed resistance") are described below (taken from BPPD 2010d; note that specific details may vary between registrations):

 EPA will receive notification within 30 days of resistance confirmation;
 Affected customers and extension agents will be notified about confirmed resistance;
 Affected customers and extension agents will be encouraged to employ alternative
CRW control measures;
 Sale and distribution of [Bt corn] in the affected area will cease immediately;
 A long-term resistance management action plan will be devised according to the characteristics of the resistance event and local agronomic needs. The details of such a plan should be approved by EPA and all appropriate stakeholders.
 
Associated Uncertainties - Remediation
 
The current generic remedial action plans may not be proactive enough to halt resistance from spreading; rather, the applied mitigation measures may suppress population densities and mask the resistance problems in the field the following year. They may actually delay the implementation of more effective measures (i.e., to reduce the spread of or eliminate resistant populations) earlier in the remediation process.  Therefore, BPPD proposes that a more proactive approach of remediation be adopted for the corn rootworm resistance monitoring program.  A primary reason for this proposal is that all currently registered toxins for corn rootworm control are considered non-high dose.  Resistance is predicted to evolve faster in non-high dose PIPs than those with high doses (Tabashnik et al. 2004), which will likely limit the time frame to successfully employ remediation measures.  Further, corn rootworm undergoes obligate diapause, which limits opportunities for timely bioassays and resistance conclusions (testing is typically done the season after field samples are taken). 
 
Species and toxin specific remedial action plans should be in place before resistance has evolved so that these plans can be deployed quickly when "unexpected Bt corn damage" is identified. Unexpected damage is likely to be a result of high corn rootworm densities and/or resistance. The earlier implementation of specific remedial action plans could reduce Agency and registrant response times by > 1 year.  BPPD further proposes that the remedial action address local (single and multiple hot spots) and widespread (large, continuous geographic areas) resistance scenarios (additional discussion in the next section). This would allow the Agency to implement remediation sooner and increase the chance of successfully containing or reversing resistance evolution in corn rootworm. 
 
The goals and objectives of the specific remedial action plans submitted should be clearly stated. With theoretical models, registrants could, for example, quantify the effects of different remediation practices on the resistance allele frequency and lifetime of the affected PIP product.  At a minimum, two remediation scenarios could be modeled:  one where resistance evolves in "hot spots" (isolated incidents) and the second where resistance evolves on a larger geographic scale.  The modeling analyses should provide qualitative insight into which remedial actions are the most mitigatory in nature.
 
Knowledge of corn rootworm's dispersal (i.e. proportion of long distance movement) will be critical for parameterizing remediation models and should, therefore, be made a major focus of research. It will also be crucial to use simulation models to quantify the proportion of dispersal that would make containment ineffective or impossible for corn rootworm. Therefore, these simulations should explore effects of different proportions of dispersal (male dispersal and pre-ovipositional female dispersal) on the success of remediation. 
  
b. Hot Spot Resistance 
 
"Hot Spot" resistance refers to incidents of resistance where the target pest evolves resistance to a pesticide in specific locations and where surrounding fields in the same geographic areas have an average resistance allele frequency well below the level detected in the hot spot. Also, the surrounding areas do not yet have decreased efficacy associated with an increased resistance allele frequency. In a hot spot, the resistance allele frequency is, therefore, significantly greater than in the surrounding area and should be visually detectable as performance problems under high population pressure/density. 
 
For Bt corn, likely causes of such resistance pockets or hot spots for corn rootworm include the following factors: continuous planting of corn-on-corn year after year associated with the use of the same Bt PIP toxins, lack of crop rotation, and/or lack of sufficient refuge. Many of these causal factors of resistance can be attributed to grower behavior. Another plausible explanation for a hot spot resistance could be long-distance dispersal of resistant adult corn rootworm into a corn field of a geographic area with susceptible corn rootworm populations.  Such infrequent long-distance dispersal of resistance genes could then be amplified and become visible as Bt corn failure during the following season when resistant offspring larvae have hatched and fed on roots of corn in the Bt field. The two scenarios described (grower behavior and pest intrinsic causes) are not mutually exclusive; hence it may be a challenge to always identify one or all factors responsible for hot spot resistance. By reviewing the management practices on the farm with the resistance hot spot, it may be plausible to exclude human behavioral causes leading to resistance and identify dispersal of resistant individuals into a field  -  but to tease apart whether that was caused by long distance or inter-field dispersal may pose another challenge. 
 
Single hot spot resistance should be easier to mitigate if pests have a low propensity to disperse but requires early resistance detection coupled with fast reaction time and effective mitigation tool(s). Corn rootworm population extirpation could be achieved by rotating to, for example, soybean. If data support that cost to Bt resistance is present in field-evolved, resistant corn rootworm populations, then planting non-Bt expressing (isoline) corn in the affected area could disfavor resistant genotypes and give susceptible genotypes a chance to increase in frequency. In this manner, the resistance allele frequency could be lowered in subsequent generations without driving the population to extinction. Additionally, it might be beneficial to plant a pyramided PIP expressing Bt toxins with different mode of action(s) to reduce population densities and further mitigate resistance through potential immigration from neighboring areas harboring susceptible populations. 
  
c.       Area Wide Resistance
 
When resistance emerges in many locations across a wide geographic area, it may be challenging to discern between cases where 1) resistance has spread out from one hot spot through dispersal (scenario 1), 2) resistance evolved in many different hot spots simultaneously from ubiquitous management practices favoring resistance (i.e., causal factors of resistance) (scenario 2), or 3) a combination of both (scenario 3). 
 
In scenario 1, if early resistance detection of a hot spot does not occur and/or implementation of mitigatory actions is delayed, then resistance genes can disperse beyond the initially affected area and radiate outward. In corn rootworm this may be a slow process because the majority of movement is short range in nature (Spencer et al. 2009, Nowatzki et al 2003b, and BPPD 2010f). The end result of this resistance scenario would eventually be widespread resistance identified by failure (unexpected damage) of Bt corn fields expressing the compromised PIP in a shared geographic area. 
 
In scenario 2, resistance evolves in multiple locations because of continuous planting of corn year after year with the same trait, lack of crop rotation, and insufficient planting of non-Bt corn. However, the end result of this resistance scenario is not likely distinguishable from the previously described scenario.  It should be perhaps expected that resistance in corn rootworm will be a combination of both scenarios. BPPD speculates though that the second scenario could be the driving mechanism for resistance in corn rootworm because of widespread agricultural practices employed across the Corn Belt (i.e., continual planting of corn for feed or ethanol production, lack of crop rotation, etc) and the non-high dose nature of the rootworm Bt toxins (see discussion in BPPD 2011 and 2012).
 
When large scale resistance becomes apparent, one mitigatory step might be to stop use of the compromised Bt toxin where crop failure has been observed and to extirpate the resistant populations by rotating to a non-host crop. This option assumes that the affected resistance zone can be accurately defined (see below).  Volunteer corn would also need to be actively removed from the remedial action zone to assist in the extirpation efforts of resistant populations.
 
Associated Uncertainties- Area Wide Resistance

Corn is grown across most regions of the Corn Belt and is largely ubiquitous.  Therefore, from the corn rootworm's perspective the landscape can be viewed as a continuous habitat with little fragmentation.  But from a resistance management perspective, this same landscape is highly fragmented because growers can choose between multiple Bt corn products from different companies. This complicates large scale mitigation for a compromised PIP toxin if resistance is identified in multiple fields scattered across a landscape but neighboring farms are not also customers of the affected registrant. 

Effective remediation requires knowledge of the geographic area that harbors the resistant population(s), which can be difficult because the failed Bt fields may not necessarily demarcate the extent of the area affected by resistance. Environmental conditions could mask unexpected root damage in Bt corn; excessive damage could be present but would not be detected because of lack of lodging (i.e., as in drought conditions). Unfavorable climatic factors (e.g., cold winter with little snow cover that increases egg mortality) can lower densities of resistant CRW populations and reduce root pruning below the unexpected damage trigger. Favorable environmental conditions (i.e., mild winters) could greatly increase densities of susceptible populations causing unexpected damage in Bt expressing fields (low and intermediate dose toxins), although BPPD contends that large corn rootworm densities (in low dose toxins Bt fields) could be indicative of resistant populations or populations in the process of becoming resistant. 

A resistance allele gradient could be used to define the remedial action area. To assure that insect samples will be available for this determination, sentinel patches may need to be maintained in or nearby to the affected fields (some distance away that is reflective of the typical adult dispersal) and areas beyond the perimeter of visible Bt failure. For this method to work, however, it will be necessary to determine what resistance allele frequency would constitute field resistance for corn rootworm. 

The allele frequency that defines field resistance might have to vary depending on the dose of toxin. When the dose of toxin is greater, the allele frequency value could be somewhat higher before a population is called resistant. When the dose is lower than high dose, the resistance allele frequency that constitutes field resistance may need to be lower to allow timely implementation of remediation (delays caused by pest biology, assays, and regulatory process). The reasoning for this proposition is based on the following:  if the high-dose refuge paradigm functions as envisioned (no significant reduction in numbers of susceptible insects due to density dependence in refuge; emergence curve overlap for refuge and Bt insects), then high numbers of susceptible insects should be available to mate with low numbers of resistant (rr-genotype) individuals.  The heterozygous resistant (and susceptible) offspring would subsequently be killed by the high dose expressing Bt PIP plant, and only rr-genotypes would survive.  The resistance allele frequency would theoretically be representative of only homozygous resistant individuals. This is not so for low dose toxins; for less than high dose Bt toxins, a portion of susceptible insects and heterozygous resistant insects can be expected to survive Bt exposure.  Once a certain resistance allele frequency has been reached in such a population, the increase in resistance allele frequency is mainly driven by surviving and more numerous heterozygous resistant individuals (a much faster process than under the high dose refuge paradigm).  On the other hand, it is important to note that this one resistance gene assumption could be complicated if multiple genes are involved with corn rootworm resistance (e.g., Lefko et al. 2008).
 



d.      Insecticide Use

One of the primary benefits of rootworm-protected Bt corn has been the potential for the reduction in chemical insecticides used to control corn rootworm (see BPPD 2010a).  BPPD is concerned, however, that these benefits will be eroded if current mitigation strategies are employed long-term and on a large landscape in response to unexpected pest damage or resistance.  To illustrate, BMPs have been implemented by Monsanto in areas where "unexpected damage" to Cry3Bb1 corn has been reported or when thresholds have been exceeded (NIS > 1.0 for single traits; NIS > 0.5 for pyramided products). Those management practices include the following (paraphrased from EPA Registration 524-606 and Monsanto's web site: http://www.monsanto.com/products/pages/crw-bmp.aspx):   
 
       Rotation to a non-host crop (e.g., soybean);
       Use of pyramided Bt corn products (e.g. SmartStax);
       Use of Cry3Bb1 corn but with additional corn rootworm control tools (e.g., soil insecticides, seed-applied insecticides, chemigation);
       Use of an alternative (non-Cry3Bb1) rootworm-active Bt corn product;
       Use of non-Bt corn with soil or foliar insecticides.

Two of the options above (continued use of Cry3Bb1 corn or non-Bt corn) involve the use of larval (soil applied) and/or adult (foliar) insecticides.  BPPD notes that soil-applied insecticide use is expected to increase in Illinois in light of reports of Bt corn resistance.  In some cases, growers may apply soil insecticides prophylactically to Bt fields in anticipation of potential resistance as "cheap insurance" (Gray 2013).  This practice has been enabled by high corn commodity prices (USDA NASS 2013), which allow growers to earn profits despite multiple control inputs (Bt corn and insecticides).

As discussed in this document, prophylactic spraying by growers and registrant BMPs to quickly deal with unexpected damage in Bt corn acres can essentially mask resistance evolution and make early resistance detection difficult. Further, the conventional insecticides used for corn rootworm control (including organophosphates and neonicotinoids) have greater risk of effects on human health and the environment than Bt corn (see BPPD 2010a). 

Rotation to a crop that does not support corn rootworm would be the most effective remediation tactic (absent volunteer corn in the rotated field) and may, depending on pest control needs, have lower environmental effects. BPPD recognizes, however, that rotating to such an alternate non-host crop may not be feasible (or desirable) for each grower because of environmental conditions (i.e., wind erosion, soil type, etc), livestock needs, land use arrangements, or other financial considerations.



VII. REFRENCES

Gassmann A., 2012c. Western corn rootworm and Bt maize: a case study illustrating the need for IPM and IRM. Presentation at the Entomological Society of America Annual Meeting.  November 13, 2012.

Gassmann A.J., J.L. Petzold-Maxwell, R.S. Keweshan, and M.W. Dunbar, 2012b. Western corn rootworm and Bt maize challenges of pest resistance in the field. GM Crops and Food, Vol. 3(3): 235-244.  http://dx.doi.org/10.4161/gmcr.20744.

Gassmann A., 2012a. Field-evolved resistance to Bt maize by western corn rootworm: predictions from the laboratory and effects in the field. J. Invertebr. Pathol., http://dx.doi.org/10.1016/j.jip.2012.04.006.

Gassmann A.J., J.L. Petzold-Maxwell, R.S. Keweshan, and M.W. Dunbar, 2011. Field-evolved resistance to Bt maize by western corn rootworm. PLOS one, Vol. 6 (7): e22629.  http://dx.doi.org/10.1371/journal.pone.0022629.

Gould, F, A. Anderson, A. Reynolds, L. Bumgarner, and W. Moar, 1995.  Selection and genetic analysis of a Heliothis virescens (Lepidoptera:  Noctuidae) strain with high levels of resistance to Bacillus thuringiensis toxins.  J. Econ. Entomol., 88 (6): 1545-1559.

Gray M., 2013. Soil insecticide use on Bt corn expected to increase this spring across much of Illinois. The Bulletin, March 28, 2013. http://bulletin.ipm.illinois.edu/?p=129.

Gray M., 2012. Continuing evolution confirmed of field resistance to Cry3Bb1in some Illinois fields by Western corn rootworm. The Bulletin, No. 20, August 24, 2012. http://bulletin.ipm.illinois.edu/article.php?id=1704.

Gray M., 2011a. Severe root damage to Bt corn observed in northwestern Illinois. The Bulletin, No. 20, August 26, 2011. http://bulletin.ipm.illinois.edu/article.php?id=1555.

Gray M., 2011b. Additional reports of severe rootworm damage to Bt corn received: questions and answers. http://bulletin.ipm.illinois.edu/article.php?id=1569 . The Bulletin, No. 22, September 23, 2011.

Lefko, S.A., T.M. Nowatzki, S.D. Thomposon, R.R. Binning, M.A. Pascual, M.L. Peters, E.J. Simbro, and B.F. Stanley, 2008. Characterizing laboratory colonies of western corn rootworm (Coleoptera: Chrysomelidae) selected for survival on maize containing event DAS-59122.  J. Appl. Entomol.  132: 189-204. http://dx.doi.org/10.1111/j.1439-0418.2008.01279.x 

Levine, E. and H. Oloumi-Sadeghi, 1991. Management of diabroticite rootworms in corn. Annu. Rev. Entomol., Vol. 36: 229-255.

Losey, J.E., L.L. Allee, V. Zbarsky, J.K. Waldron, and E.J. Shields, 2003.  Transect sampling to enhance efficiency of corn rootworm (Coleoptera: Chrysomelidae) monitoring in corn.  J. Econ. Entomol., 96 (5): 1420-1425.

Meihls, L., 2010.  Development and characterization of resistance to transgenic corn in western corn rootworm.  Ph.D. thesis submitted to the graduate school at the University of Missouri-Columbia.

Meinke L., T. Hunt, G. Kruger, R. Wright, D. Wangila, and K. Miwa, 2012.  Evaluation of western corn rootworm susceptibility to rootworm Bt traits in Nebraska.  Presentation at the Entomological Society of America Annual Meeting.  November 13, 2012.

Nowatzki, T.M., S.A. Lefko, R.R. Binning, S.D. Thompson, T.A. Spencer, and B.D. Siegfried, 2008. Validation of a novel resistance monitoring technique for corn rootworm (Coleoptera: Chrysomelida) and event DAS-59122-7 maize.  J. Appl. Entomol. 132: 177-188.  http://dx.doi.org/10.1111/j.1439-0418.2008.01270.x 

Nowatzki T.M., B.D. Siegfried, and L.J. Meinke, 2003a. Comparative movement and mating behavior of adult western corn rootworm (Coleoptera: Chrysomelidae) in a YieldGard rootworm transgenic and conventional corn field. Poster presentation at the Entomological Society of America meeting in Cincinnati, OH.

Nowatzki, T.J., N. Bradley, K.K. Warren, S. Putnam, L.J. Meinke, D.C. Gosselin, F.E. Harvey, T.E. Hunt, and B. Siegfried, 2003b. In-field labeling of western corn rootworm adults (Coleoptera: Chrysomelidae) with Rubidium. J. Econ. Entomol., 96 (6): 1750-1759.

Oleson, J.D., Y. Park, T.M. Nowatzki, and J.J. Tollefson, 2005. Node-injury scale to evaluate root injury by corn rootworms (Coleoptera: Chrysomelidae).  J. Econ. Entomol., 98 (1): 1-8.

Oswald, K.J., B.W. French, C. Nielson, and M. Bagley, 2011.  Selection for Cry3Bb1 resistance in a genetically diverse population of nondiapausing western corn rootworm (Coleoptera: Chrysomelidae).  J. Econ. Entomol., 104 (3): 1038-1044.

Roush R.T. and G.L. Miller, 1986. Considerations for design of insecticide resistance monitoring programs. J. Environ. Ent., 79(2): 293-298.

Spencer J.L., B.E. Hibbard, J. Moeser, and D.W. Onstad. 2009. Behaviour and ecology of the western corn rootworm (Diabrotica virgifera virgifera LeConte). Agricultural and Forest Entomology, 11: 9-27.

Tabashnik, B.E., 1994. Evolution of resistance to Bacillus thuringiensis. Annu. Rev. Entomol., 39: 47-79.

Tabashnik B.E., F. Gould, and Y. Carriere, 2004. Delaying evolution of insect resistance to transgenic crops by decreasing dominance and heritability. J. Evol. Biol., 17(4):  904-912; discussion 913-918.

USDA NAS, 2013. Agricultural Prices. ISSN:1937-4216. http://www.usda.gov/nass/PUBS/TODAYRPT/agpr0713.pdf

Zukoff, S.N. and B.E. Hibbard, 2012.  Evaluation of the potential development of cross resistance between eight transgenic corn types in western corn rootworm (Diabrotica virgifera virgifera).  Presentation at the Entomological Society of America Annual Meeting.  November 13, 2012.

EPA References:

BPPD, 2013a. Review of Corn Rootworm Resistance Monitoring Submissions and Data for Bt corn products containing Cry34/35Ab1. Memorandum from A. Reynolds to A. Sibold, dated March 20, 2013.

BPPD, 2013b. Review of Corn Rootworm Resistance Monitoring Submissions and Data for Bt corn products containing mCry3A.  Memorandum from A. Reynolds to A. Sibold, dated May 15, 2013.

BPPD, 2013c. BPPD IRM Team review of Monsanto's 2011 corn rootworm monitoring data and unexpected damage reports for Cry3Bb1 expressing Bt corn.  Memorandum from J.C. Martinez and A. Reynolds to A. Sibold, dated September 26, 2013.

BPPD, 2012. BPPD IRM Team review of Monsanto's 2010 corn rootworm monitoring data, unexpected damage reports for Cry3Bb1 expressing Bt Corn and academic reports of Cry3Bb1 field failures as well as corn rootworm resistance. Memorandum from J.C. Martinez and A. Reynolds to A. Sibold, dated October 11, 2012.

BPPD, 2011. BPPD Review of Reports of Unexpected Cry3Bb1 Damage, Monsanto's 2009 Corn Rootworm Monitoring Report, and Revised Corn Rootworm Resistance Monitoring Plan for MON 88017, MON 88017 x MON 810, MON 863, MON 863 x MON810, MON 89034 x TC1507 x MON 88017 x DAS-59122-7, and MON 89034 x MON 88017. Memorandum from J.C. Martinez to M. Mendelsohn, dated November 22, 2011.

BPPD, 2010a. Biopesticide Registration Action Document for Bacillus thuringiensis Cry3Bb1 Protein and the Genetic Material Necessary for Its Production (Vector PV-ZMIR13L) in MON 863 Corn (OECD Unique Identifier: MON-ØØ863-5). http://www.epa.gov/oppbppd1/biopesticides/pips/cry3bb1-brad.pdf

BPPD, 2010b.  Bacillus thuringiensis Cry34Ab1 and Cry35Ab1 Proteins and the Genetic Material Necessary for their Production in Event DAS-59122-7 Corn Biopesticide Registration Action Document (BRAD).  Revised September, 2010.  Available at http://www.epa.gov/oppbppd1/biopesticides/ingredients/tech_docs/brad_006490.pdf.

BPPD, 2010c.  Biopesticide Registration Action Document:  Modified Cry3A Protein and the Genetic Material Necessary for their Production (Via Elements of pZM26) in Event MIR604 Corn SYN-IR604-8.  Revised September, 2010.  Available at http://www.epa.gov/pesticides/biopesticides/pips/mcry3a-brad.pdf.

BPPD, 2010d. Terms and Conditions for Bt Corn Registrations.  Available at http://www.epa.gov/oppbppd1/biopesticides/pips/bt-corn-terms-conditions.pdf.

BPPD, 2010e. Biopesticide Registration Action Document:  Cry1Ab and Cry1F Bacillus thuringiensis (Bt) Corn Plant-Incorporated Protectants.  Updated September, 2010.  Available at http://www.epa.gov/pesticides/biopesticides/pips/cry1f-cry1ab-brad.pdf.

BPPD, 2010f. Science Assessment of Pioneer's modeling addenda in support of their application of registration for Optimum AcreMax1 Insect Protection. Memorandum from J.C. Martinez to M. Mendelsohn, dated February 4, 2010.

EPA Scientific Advisory Panel (SAP), 1998.  Transmittal of the final report of the FIFRA Scientific Advisory Panel Subpanel on Bacillus thuringiensis (Bt) Plant-Pesticides and Resistance Management, Meeting held on February 9-10, 1998.   Report dated, April 28, 1998.  (Docket Number:  OPPTS-00231).

EPA Scientific Advisory Panel (SAP), 2002. Transmittal of meeting minutes of the FIFRA Scientific Advisory Panel Meeting held August 27-29, 2002 on corn rootworm Plant-incorporated Protectant non-target insects and insect resistance management issues. Memorandum from Paul I. Lewis (DFO) to Marcia E. Mulkey (Director of OPP) on November 6, 2002.

