Agrichemical giant Bayer helped fund a study by university academics, then pressured them to omit photos that implicated a defective insecticide-treated seed product as a threat to bees, according to communications obtained by U.S. Right to Know.
Several seed and insecticide companies, including Bayer, paid Ohio State University researchers to determine how much their insecticide-coated seed products affected bees during corn planting season in 2014 and 2015. After the researchers presented their preliminary results to “stakeholders,” which included funders, a Bayer official asked that their final report exclude photos of insecticide-coated corn seeds in which the product appeared defective. He also urged the researchers to qualify statements in the final report that discussed threats to bee health in ways that benefited Bayer’s corporate interests.
Although the photos and all the researchers’ conclusions ultimately made it to the final publication, internal emails show the seed and chemical industry funders intensely scrutinizing the researchers’ findings during pre-publication presentations. The study’s funding contract allowed funders to review and comment on findings prior to publication, and required pre-approval for any press releases or sharing of results.
This situation serves as an example of how agrichemical companies attempted to influence scientific research at a public university. Emails show how an industry funder tried to control and spin researchers’ results. Internal communications and research contracts are elements of sponsored research that are typically hidden from the public, but in this case, they provide insight into corporate sponsors’ involvement in the research process.
Concerns about bee deaths
The main funders of this research were chemical companies that manufacture neonicotinoids, the world’s most widely-used insecticide. Neonicotinoids, or neonics for short, are often delivered to crops via a colorful coating on the seed. Neonics are systemic, which means the seed grows into a plant containing the insecticide throughout, so it kills any bug that bites it. Since they are insecticides, neonics are also toxic to bees and other beneficial pollinator insects.
The companies manufacturing neonics have been trying to shape the narrative about bees and insecticides for more than a decade, as concerns have grown about harm to bees.
In 2008, there was a major honey bee die-off in Germany that researchers traced back to the planting of neonic-coated seeds. The coatings were flaking off the seeds as tractors drove around in the fields towing corn planters, which are mechanical devices that plant corn seeds one at a time into the ground. The mechanical planting process churned up clouds of neonic-imbued dust that killed bees. Bayer said it was a bad batch of seed, and they released a paper the following year announcing they would retrain seed manufacturers, improve the coating ingredients, and develop dust-cutting modifications for planters. But a 2012 Purdue University paper listed neonic-contaminated dust from corn planters as a route of neonic exposure for bees. Canada was able to link bee deaths to corn and soybean planting events in 2012 and 2013.
The next year, the nonprofit Pollinator Partnership formed the Crop Dust Research Consortium (CDRC), which funded research to determine the best practices for minimizing honey bee exposure to seed dust. Its funders included companies such as Bayer, Syngenta, and BASF, all of which manufacture insecticidal seed coatings; trade groups with direct financial interests in neonic-coated seeds; and beekeeping and agricultural interest groups.
The CDRC said that its goal across all research objectives was to “produce peer-reviewed published papers to advance the understanding of the issue through open and transparent oversight.” Its material explicitly said the research was not intended as an endorsement of seed treatment, neonics, or any other specific practice. One of its objectives was to test a new lubricant Bayer developed that was intended to contain the coatings better than existing lubricants.
In 2014, the CDRC granted $157,224 to Dr. Reed Johnson and agronomist Harold Watters of Ohio State University. The researchers were to characterize bee visits to flowers around cornfields during spring planting to figure out how to reduce pesticide exposure during those visits, and to test how well various seed lubricants reduced dust. In 2015, CDRC granted the lab an additional $145,000 to continue the research. In addition, the researchers were to investigate the long-term health consequences of bee colonies exposed to insecticide-contaminated dust from corn planters, and the efficacy of the CDRC’s recommendations in preventing bee exposure to planter dust. Researchers in Iowa, Nebraska, Montana, and Ontario also received funding.
Johnson and Watters determined there was very little difference in the amount of neonic dust generated by different kinds of corn planters, and that Bayer’s new lubricant didn’t outperform the other lubricants. This conclusion was published in the CDRC final report in 2017. Johnson and Watters also documented higher neonic levels in bee-collected pollen, and an uptick in bee deaths during corn planting season. However, they didn’t find evidence that the contaminated pollen killed the bees (bees could have died by direct exposure to the planter dust) and the neonic exposures didn’t appear to impact longer-term colony strength or overwintering success. They published these conclusions in Environmental Toxicology and Chemistry in 2020, in addition to the CDRC final report.
The conversations leading up to publication show that, although the researchers ultimately published all their findings, they first had to respond to funders who had a financial interest in the results as they reviewed and suggested changes to their report.
The flaky seed coating ordeal
During corn planting, Johnson and Watters noticed the insecticidal coating on their seeds was visibly flaking off, so they sampled the seeds after they’d been rattling around in the planting hopper for 15 minutes to an hour. They later photographed a few seeds from every sample. The photos showed the seeds were losing a lot of their insecticidal coating before they went into the ground. Johnson and Watters included these photos in a November 2015 presentation to the Entomological Society of America and a contractually-required December 2015 presentation to CDRC stakeholders.
By January, the CDRC’s industry-affiliated stakeholders began to take intense interest. In emails from January through November of 2016, they asked Watters and Johnson for minute details about their research conditions, remarked that the seeds looked unusually bad, and suggested conducting more studies to replicate the phenomenon.
“I can’t explain the photos taken by OSU,” wrote David Fischer, the director of pollinator science for Bayer, prior to a July 2016 presentation of the Ohio State team’s preliminary results. “I’d like to know more about the seed treatment quality. … So much of the seed treatment has been eroded from these seeds I don’t see how any pest control efficacy could be achieved. In past studies, the amount of active ingredient removed from the seed is usually less than 2%. The seeds from the OSU study look to be 30% or more.”
While Fischer insisted the findings were an outlier, Watters defended his research and said the coating problems he observed were likely typical in the field, in an interview with U.S. Right to Know.
One important distinction is that most of the prior experiments Fischer referenced were conducted on small plots, Watters said.
“Frequently when we do corn trials in small plots we use a length of 17 feet 5 inches – because in a 30-inch row this is 1/1,000th of an acre. It makes calculations for yield very easy. I do not use this for most of my work because I feel it is too small to get world/reality check as to accuracy,” he said. “A typical farm field is 40, 80, 160 or perhaps 320 acres in size. My work was done in these typically sized farm fields.”
Furthermore, Johnson and Watters’ research was conducted on working farmers’ fields, using the coated seed the grower already planned to use for that field.
“The seed came off the typical commercial lots sold to farmers,” Watters said. “So hybrid, company of origin, maturity, etc. would have been at random – we made no prior plans on seed, only what field we were in and with what grower.”
Ultimately, Watters said they observed degraded seed coatings at random throughout the study.
“Our observations were totally random and as a conclusion I would expect similar seed coat breakage to happen everywhere,” he said.
At least one stakeholder on the conference calls told Watters and Johnson that getting seed treatments to stick to corn seeds is difficult as a rule, Watters said.
“This was apparently not unexpected for at least that one person.”
“I think there was an understanding that this is probably a component of the problem: the quality of the coating maybe was not as good as the seed casing they were producing in their own facilities,” Johnson said, referencing an American Seed Trade Association top lobbyist asking about how the seeds had been stored and under what circumstances they’d been purchased. But he said he believed the scrutiny was coming from industry affiliates’ interest in identifying what to fix. “I think they genuinely wanted to solve this problem,” he said.
“The recollection I have is they couldn’t believe that there were actually flakes of the seed coating coming off of the seed,” Watters said when asked about the emailed exchanges. “The seed coats did crack. We collected material in the field, we had sticky traps… there were chunks of the seed coat that were coming off as the seed was being planted.”
Although he engaged with the questions from the industry stakeholders, Watters said he remained firm in his testimony.
“I’m comfortable with the remarks I made with them that these were the results I saw,” he said.
Watters also found a 2013 review from Belgium that documented similar seed coat loss. He wrote to Johnson in August 2016 that “apparently it is not unheard of… and should not be a surprise.”
Bayer’s David Fischer asks for changes in the report
In advance of a conference call with other research stakeholders in July 2016, Bayer’s Fischer wrote to the group:
“Bayer has very significant concerns about the inclusion in the CDRC report of the photographs of seeds with severely eroded seed coatings that are shown in Fig 10 of the report from Ohio State University, especially the photo characterized as “typical” example of what seeds look like after passing through a planter. Many BCS [Bayer Crop Science] personnel have seen these photos and consider them completely atypical of anything they’ve seen in their experience.”
Fischer went on to say that Bayer’s internal studies yielded very different results, where the coating stayed on nearly perfectly after passing through a planter. Fischer then asked for further studies and interpretations on the matter.
The CDRC final report indicates which seed companies sold the seeds used in Johnson and Watters’ study (Beck’s, Dekalb/Monsanto, Master’s Choice and Stewart’s) but no information explicitly named which seed coating manufacturer’s product was on those seeds. Although three seed coating manufacturers were represented in the CDRC’s stakeholder group (Syngenta, BASF, and Bayer), Bayer’s David Fischer dominates the communications obtained by USRTK.
After the meeting, Fischer sent Johnson several pages of comments in which he asked for Johnson to remove the photos, and asked him to qualify several statements pertaining to bee health.
Despite this, conclusions remained unchanged between the June 2016 draft that Fischer commented on and the final report, aside from one statement that Johnson added to qualify the observed uptick in bee deaths: “The increase in adult mortality, however, would not be expected based on the concentrations of seed treatment insecticide measured in bulk pollen samples.”
In other words, it clarified that pesticide concentrations in the pollen were too low to adequately explain the bee deaths they observed.
“The insecticide was probably not evenly distributed throughout the pollen,” Johnson explained, referring to it as “the toxic chocolate chip cookie hypothesis.”
“There were these little chunks of the seed treatments and if the bee got that chunk, it was gonna die. But overall, the concentration, if you averaged it all together, was still below the level that was expected to kill bees,” he said in an interview with U.S. Right to Know.
Although Fisher’s requests appeared to elicit little change in the report, he asked for many changes: With regards to the flaky seed coatings, he called for a dedicated study to determine whether OSU’s seeds met industry standards, but noted that might be impossible for the specific batch of seeds that had been used in Johnson’s study.
“So the quality of the seed treatment of the seeds used in the OSU study is an uncertainty. Given this uncertainty, I would suggest that the photographs of abraded seeds be removed from the report, or at least a clear statement made that indicates the quality of the treatment of these seeds is unknown,” Fischer wrote. “Figure 10, showing the photos of seeds with a large amount of the seed treatment eroded as a result of planter abrasion is problematic. There are no data backing up the statement that the photo on the left represents a typical seed after passing through a planter. Also, the amount of a.i. [active ingredient] remaining on the seed hasn’t been determined. These photos have the potential to mislead the reader regarding the amount of a.i. lost due to dust abrasion. These photos should be deleted from the report.”
Fischer further suggested qualifiers and changes to the report, such as removing the word “elevated” from “[neonic] residues are reliably detected at elevated levels (8 ppb above background on average).” This suggestion was ignored.
Fischer also suggested that Johnson contextualize the neonicotinoid levels he found in pollen that bees had collected by discussing their risk to bee health. He provided several paragraphs of analysis and a diagram to that extent.
“The residue levels measured DO NOT indicate any appreciable risk to honey bees, and DO NOT explain the acute mortality observed in the study. The above analysis should be included in the report. Without it, the reader might mistakenly think that the residue measurements DO explain the mortality observed.” (Emphasis Fischer’s.) This appears to be the one area where Johnson et al added a sentence to emphasize the above fact, although Fischer’s diagram was not included in the final report.
Fischer also wrote to Johnson about the bee deaths documented in the research.
“It is important that the CDRC report emphasizes that the level of mortality documented in this study was low, and had no observed effect on colony development or viability.” he said.
Fischer went on to tell Johnson that one of his conclusions, that removing flowering weeds is unlikely to reduce pesticide exposure to bees, “needs to be considered speculative, because the risk to bees was very low regardless of the landscape conditions in this study.” This was ignored in Johnson’s section of the final report.
Fischer also told Johnson that “the number of trials is too few to draw the definitive conclusions” his lab made about the relative efficacy of Bayer’s seed planting lubricant.
“Clearly, the photos of seeds have struck a nerve. I’d be very reluctant to remove the photos entirely, but I think David Fischer does have valid concerns,” he wrote.
When asked in an interview about what Fischer asked him to do, Johnson said he thought Fischer didn’t want the pictures to be highlighted in the report.
“I pushed back,” Johnson said, because he believed the degraded seed coatings could be an important factor in the questions the CDRC sought to answer. However, Johnson acknowledged that he didn’t know much about coated seeds, and Fischer’s concerns seemed valid if the seeds he’d photographed had been a fluke. He largely disregarded Fischer’s suggested edits, though, besides adding the line about how the average insecticide levels were lower than lethal in the pollen samples.
“I think we were happy with the conclusions we had come to. I guess we disagreed with what David Fischer had suggested,” he said, “except on that one point.”
Did Johnson think this exchange was an overstep on Fischer’s part?
“I think that was his role to play in this group, and he did it,” Johnson said.
When asked whether he thought several stakeholders’ commercial interests in coated seeds affected the discussion of his study results, Johnson said he thought their presence was productive.
“It was interesting because they definitely came at this with their point of view, which was expected,” he said.
They were engaged on the conference calls and there was sometimes genuine disagreement, he said, about things like the efficacy of the lubricants, the mechanism of bee exposure to neonics, and whether weed control was a good approach to mediate the effects of seed treatment dust.
But the group was well-mediated, and it was “really exciting” to be on calls with the very people who had the ability to make changes and solve the problem, he said.
“In many ways they’re the prime audience for this research. I felt the discussions were quite productive. I always had it in my mind they were coming at it with their point of view,” he said, adding that he freely presented the study’s results to many audiences, and the chapter he submitted to the CDRC final report was published with all data and conclusions intact.
Funder/researcher interactions consistent with funding contracts
The funding contracts Johnson signed in 2014 and 2015 specify an obligation to share results with the funders before publication. “The Corn Dust Research Consortium has been formed to review proposals and oversee the project execution, including review and comment on study protocols, draft reports and presentation materials prior to their execution and public release. Final decisions on technical interpretation of the study findings and content of study reports, publications and presentations will be made by study personnel,” read the CDRC request for proposals. Both contracts Johnson signed contained provisions that all press releases and release of research results must be pre-approved by the Pollinator Partnership. Furthermore, “Research data is not to be released outside of the CDRC group without an agreed upon date between P2 [Pollinator Partnership] and the principal investigator,” the contracts read, with a suggested date soon after that of each contract’s expiration.
The contracts explicitly encouraged “photography or videography documentation from the onset of methods, experimental plots locations and in field activity for use in explaining the methodology and illustrating observations.”
“That was one of the expectations going into this project… they said they would let us publish whatever we wanted,” Johnson said. When asked about the contract’s pre-approval clause, he said that only factored in during the length of the contract. “They were actively encouraging us to publish our results,” he said.
Dr. Quinn Grundy is an assistant professor at the University of Toronto, who studies corporate influence in health and health research. She said elements of this situation at Ohio State are “really familiar” to what she sees in the medically-related industries.
Overall, the CDRC contract “sounds pretty standard” to Grundy: “I think, in most cases, sponsors ask for updates or the ability to comment or to approve publication,” she said.
However, the clause in the CDRC contract requiring pre-approval of press releases isn’t ideal: “It definitely raises red flags for me, and I think in an ideal world the sponsor would have no role in approving the output of sponsored research, but unfortunately I think it’s pretty common. Usually the clause is something like the sponsor has the right to see, but can’t block publication or overrule the researchers, in the decision to publish,” she said.
It’s “so important” that contracts explicitly state a researcher’s ability to publish their research regardless of the results, Grundy said, particularly for early career scientists.
Grundy also noted that the CDRC is one instance of many where corporations have obscured their research contributions by providing them through a foundation or trade organization. Such entities might have “innocuous and nice sounding names,” Grundy said, but are “actually representing the pooled contributions of several commercial or corporate entities with an interest in shaping the direction of a research agenda.”
It can be hard for an average reader to identify conflicts of interest in a system that relies on voluntary disclosures from researchers, Grundy said. Structural or policy interventions might be necessary to eliminate bias or funder influence. “It sounds like [Johnson and Watters] in fact did the right thing, but could have benefited from more supports to ensure the independence and integrity of their work which… has some pretty powerful implications for not only bees, but all of us that depend on bees,” she said.
One beekeeper stakeholder’s perspective on chemical industry influence
Iowa Beekeeper Manley Bigalk was appointed a CDRC stakeholder on behalf of the American Beekeeping Federation, a beekeeping industry interest group. He said that the chemical company stakeholders contributed more funds and had a more central role in discussions.
“So I was a stakeholder for, I forget what was pledged, $300 or $700 or something like that. Well, the companies were in there for hundreds of thousands because they were sponsoring the tests,” Bigalk said.
“A chemical company running grants to find out if their chemicals are active? Might be a little bit of a conflict of interest there. But a lot of the research today is funded by these people even at the universities,” he said. “And they have the money.” Bigalk spoke about the way he saw the main sponsors, the chemical company stakeholders, attempt to shape the scope of the study and the content of the report. “There never was a full admission there was dust in the air, to tell you the truth,” he said. Bigalk also noted they left certain critical details out of the CDRC’s final report. “Wouldn’t you think,” he said, if the study sought to address beekeeper complaints about toxic corn dust, “ …that you would have some evidence of things that now have been corrected? Or evidence that it does kill bees? Wouldn’t you think?”
Although Bigalk saw the project through from start to finish, another beekeeping interest group representative grew so disheartened with the CDRC that he dropped out of the partnership. Randy Verhoek attended a couple meetings on behalf of the American Honey Producers Association, but dropped out after hearing the group arrive at conclusions he didn’t agree with, American Honey Producers Association president Chris Hiatt confirmed.
“It was really a strange experience. And for more ways than one,” Bigalk said. In one of the very first stakeholder meetings, Bigalk said he suggested the seed coating dust could be remediated with a different polymer that would stick better to corn seed. “That comment was made and there was no response and they went on to something else,” he said. That surprised him, he said, since a small chemical vendor had readily mentioned the coating problem to Bigalk previously.
Johnson’s team stands their ground
Several days after Johnson responded to Fischer’s feedback, Pollinator Partnership personnel scheduled a phone conversation with Johnson, then convened another conference call in November specifically to talk about the seed coating issue.
During that call, minutes obtained by U.S. Right to Know show that Fischer said that “chunks coming off seeds do not affect the bees as much as the pollen-sized seed coating that comes off,” to which then-Pollinator Partnership Executive Director Laurie Davies Adams responded that big chunks of insecticidal coating could still affect the water quality on the site.
Meeting minutes show one seed company representative, Beck’s Hybrids manager Jim Herr, suggested throwing out the worst-looking photo. “In trying to understand how this could have happened, it seems that this one photo (picture 2. B-F) was an extreme scenario so it may not be the best to include in a report.”
The CDRC final report included 14 photos of the degraded seeds, including the contested “extreme” photo mentioned above.
“All examined corn seeds showed signs of seed treatment degradation, varying in severity,” the report reads. The photo caption says that “seeds varied in the perceived integrity of the coating, but most showed signs [of] particle generation. In an extreme case, we found one seed that had almost no coating left on it (B2), though this was the only example we noticed of such extensive degradation.”
In an early November email exchange, Watters and Johnson expressed their “exhaustion” on the topic to one another.
“However I get the sense there may be some appetite for funding additional research? Hard to say,” Johnson wrote.
Ultimately, Johnson said that he didn’t conduct any of the further studies that had been suggested. “I would really have loved to follow up,” he said.
He described a study he wanted to do where he’d use a dust meter to measure the amount of coating flaking off of the seeds, but it was “actually quite difficult” to access the necessary equipment, the CDRC didn’t offer any further funding opportunities, and eventually, Johnson moved on.
“I’m a honey bee researcher, and I finally made the decision that this is really getting too far afield,” he said.
Other issues with the final CDRC report
Starting in January 2017, the Pollinator Partnership circulated a draft of their final report, which was meant to summarize the study results and make final recommendations for mitigating planter dust contamination. Members of Johnson’s lab internally expressed disdain for the long list of recommendations to farmers, beekeepers, pesticide and equipment manufacturers, seed dealers, regulators, and educators. They were particularly irked that the report made recommendations to beekeepers, whose practices were not the focus of the study.
“I’m genuinely offended that the farmers and beekeepers are the first groups to be addressed, as though they bear the primary responsibility for this issue,” a lab member whose name was redacted told Johnson when he asked them what they thought of the conclusions.
When submitting comments on the report’s final draft in April, Johnson wrote, “I don’t recall anyone doing research on supplemental feeding or providing water or placement of hives around a field. I realize this leaves you with no bolded recommendations for beekeepers, but generating recommendations for beekeepers really wasn’t a focus of the research,” he wrote.
Nevertheless, the CDRC’s final report featured bolded recommendations for beekeepers.
Furthermore, the report’s overall conclusions recommended that farmers remove flowering vegetation within fields “through tillage, mowing or use of herbicides where appropriate” prior to planting, despite Johnson’s chapter recommending against that. Johnson said he believes some of the above discrepancies came from the fact that, although the 2017 report only contained the final year of each university team’s research data, the report’s overall conclusions attempted to synthesize three years of sometimes conflicting conclusions made by the different university research groups.
Johnson believes that Pollinator Partnership staff drafted the overall conclusions. Pollinator Partnership did not respond to a request to confirm this.
“My conclusion is that, at the end of this project our major conclusion was that it was really a quality control issue… and the seed treatment wasn’t adhering well to the seed,” Johnson said.
He said he believes that the seed coating formulation changed in the years since his study.
“I don’t get any calls about bees dying during corn planting anymore,” he said, compared to many calls in 2012 and 2013, and he detects lower insecticide concentrations in pollen during planting season nowadays. “It seems that this problem has been solved one way or another,” he said. “I think [the research] reached an audience who needed to see it.”
“I still don’t…like these insecticidal seed treatments, for a whole variety of reasons,” Johnson said. “The bee issues are just one of them. The fact that they’re not really very effective at pest control is another major issue here. But they do seem to have solved it causing mass bee kills from dust off like it did in these years.”
The CDRC research did not signify the end of neonic contamination issues, however, Bigalk said.
“They made the steps to research the corn dust situation. But in many ways, it just stopped short of showing the full toxic effect of neonics on what we call beneficial insects,” he said, or on aquatic organisms.
Bigalk said he believes the participating chemical companies corrected the flaky seed coating formula, as Johnson suspects. He also believes that mechanical dust control kits, which farmers can attach to corn planters, have helped to reduce dust since the CDRC was formed.
“We have controlled a great amount of that dust,” he said.
Although Bigalk had a large bee kill prior to the formation of the CDRC, he hasn’t had anything like it in the years since, he said. But he said problems from the “abuse of chemicals” persist.
“But with the things affecting honey bees today during the summer months and into the winter, we don’t have outright kills…. we’ve got this invisible negative influence that are coming into the hive and are just disrupting many, many things. And it’s a subtle thing,” Bigalk said, which might manifest in a colony’s queen dying or never returning to the hive after going out to mate. “My word, there’s such a dramatic change in colony health today compared to what it was 30 years ago. 40 years ago,” he said. Although parasitic Varroa and tracheal mites hurt colony health, Bigalk said they’re not the only bee stressor at play. “There are beekeepers out there without any elevated number of Varroa mites that are suffering greatly. It’s invisible…. Invisible negative influences from abuse of chemicals,” he said.
Just last year, Nebraska regulators closed down the AltEn LLC ethanol plant after the neonic coatings on the seeds the plant processed caused widespread environmental contamination.
Bigalk is a farmer as well as a beekeeper. He said he’s aware that neonicotinoids are applied to far more acres than actually need treatment.
“It’s abusive use,” he said.
However, he still allows them to be applied on his own land, in part to make things less complicated for his neighbors, since their fields are planted with the same equipment.
“I would surely prefer not having it on, but it doesn’t work out that well when you’re in a custom [planting] situation,” he said.
The Pollinator Partnership did not respond to requests for comment on this story. A Bayer representative informed USRTK that David Fischer has retired. The representative did not respond to requests for comment.
U.S. Right to Know obtained the documents for this report through an Ohio Public Records Act request to Ohio State University, and by asking Dr. Johnson directly for certain additional communications.
Abbe Hamilton is an investigative reporter covering corporate influence on neonicotinoid science and policy for U.S. Right to Know.