lnt gate rebuttal (Edward Calabrese)

Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Environmental Research journal homepage: www.elsevier.com/locate/envres

Review article

LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment Edward J. Calabrese Department of Environmental Health Sciences, Morrill I, N344, University of Massachusetts, Amherst, MA 01003, United States

art ic l e i nf o

a b s t r a c t

Article history: Received 29 March 2016 Accepted 30 March 2016

This paper provides a detailed rebuttal to the letter of Beyea (2016) which offered a series of alternative interpretations to those offered in my article in Environmental Research (Calabrese, 2015a) concerning the role of the U.S. National Academy of Sciences (NAS) Biological Effects of Atomic Radiation (BEAR) I Committee Genetics Panel in the adoption of the linear dose response model for cancer risk assessment.

Significant newly uncovered evidence is presented which supports and extends the findings of Calabrese (2015a), reaffirming the conclusion that the Genetics Panel should be evaluated for scientific misconduct for deliberate misrepresentation of the research record in order to enhance an ideological agenda. This critique documents numerous factual errors along with extensive and deliberate filtering of information in the Beyea letter (2016) that leads to consistently incorrect conclusions and an invalid general perspective. & 2016 Elsevier Inc. All rights reserved.

Keywords: Linear dose response Cancer risk assessment Dose response Hormesis Mutation Threshold dose response

1. Introduction Beyea (2016) offers a series of alternative interpretations to my conclusions (Calabrese, 2015a) on the acceptance of the LNT by the regulatory and public health communities and the role of the radiation genetics community and the U.S. NAS BEAR I Committee, Genetics Panel in this process. The principal claims of Beyea are that the BEAR I Committee, Genetics Panel was not biased, despite being portrayed this way in my series of papers (Calabrese, 2013b, 2014a, 2015a, 2015b, 2015c), that the Genetics Panel was not particularly influential to subsequent NAS committees dealing with the issue of the effects of low doses to mutagens and carcinogens, especially the BEIR committees, and that LNT would have emerged as the default assumption in cancer risk assessment even if BEAR I Genetics Panel had not been convened. In making these claims he creates his own narrative, a mixture of filtered information with the end result resembling a type of historical novel with a blend of facts, fiction, numerous factual errors, revisionist history and wishful thinking. Let us now examine the scholarship and accuracy of Beyea's claims and how his story holds together.

2. The Genetics Panel: biased or not The Genetics Panel recommended that the dose response DOI of original article: http://dx.doi.org/10.1016/j.envres.2016.01.039 E-mail address: edwardc@schoolph.umass.edu

model for genetic risk assessment be switched from a threshold to a linear model in 1956 without discussion or debate based on meeting transcripts. During the second meeting of the Panel (February 5–6, 1956) in Chicago, Illinois panelist Dr. Tracey Sonneborn, University of Indiana, read into the record what amounted to a “Geneticists’ Creed”. This mantra may be summarized as: the dose response for ionizing radiation induced mutations was linear, even down to a single ionization, and that all mutagenic damage was cumulative and irreversible. Despite a Panel of 17 outstanding researchers and leaders, including 13 nationally recognized experts in radiation genetics, no dissent was offered. This lack of discussion/debate was unexpected as I was hoping to learn how leaders of the past wrestled with the key issue of the nature of the dose response in the low dose zone. I wanted to learn who/what influenced the debate, to see the ebb and flow of the intellectual exchanges and what perspectives were persuasive. However, the meeting resembled an assembly of committed believers to an ideology, where discussion, debate as well as demands for evidence on the crucial question were not part of the agenda. The Panel would continue to be guided by this Geneticists’ Creed throughout their entire course of activities. Their foremost goal was not to address how genetic risks from exposure to ionizing radiation should be assessed, but how to implement their dose response mantra into policy and risk assessment practices. This specific group of radiation geneticists was convened to create a dose response revolution and this process started with decisions concerning who would comprise the Panel.

Creating a Panel that uniformly agreed on the central issue of a highly contentious debate was biased from the start, denying a fair

http://dx.doi.org/10.1016/j.envres.2016.03.040 0013-9351/& 2016 Elsevier Inc. All rights reserved.

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

2

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

and legitimate process to properly assess what is the nature of the dose response in the low dose zone. The end result was a selffulfilling prophesy. Selection of individuals with a uniform belief core on this issue precluded the search for truth. That this situation would occur should not have been a surprise as each Panelist's beliefs on the issue were well known.

Some of the Panelists had been/were being funded by the Rockefeller Foundation, the same organization that was also funding the BEAR Committee activities and selecting the members.

The President of the Rockefeller Institute for Medical Sciences (later to become Rockefeller University) and the President of the NAS at this time was the same person, Dr. Detlev Bronk. President Bronk would ensure the creation of a Genetics Panel and that the person (i.e. Dr. Warren Weaver) overseeing the funding of the geneticists’ grants from the Rockefeller Foundation would chair the Genetics Panel. Thus, organizational and leadership structure of BEAR I, as well as the Genetics Panel selection and functioning, displayed an improper inbreeding relationship between the funder and recipient organization with the potential for undue influence of the funding organization over Panelists who were highly dependent on grant funding from the Rockefeller Foundation. From the start, the process would be susceptible to manipulation, coercion and bias, as well as timing pressures. While the complexity of these inter-relationships makes their dissection extremely difficult, too much power, control and influence was rested in Bronk and Weaver. This type of organizational framework, functional control and funding manipulation, created the likelihood that the funder would get what it wants.

While the bias of the Panel started with its selection, it was reaffirmed and transformed into behavior/actions and policy recommendations. The record shows that alternative dose response models were ignored. In the one instance when a threshold model could have been discussed the presentation was inaccurate with low dose data being dropped [e.g. see Calabrese (2013b) for a discussion of the presentation of Panelist Bernard Kaufmann who misrepresented/omitted data supporting a threshold response for radiation induced mutation]. Such leadership failures are also seen in a decision (i.e. the Panel voted) (Weaver memo to Panel, June 5, 1956d (Weaver, 1956d)) to hide profound disagreements of the Panelists over the estimation of radiation induced populationbased transgenerational genetic damage risks in the U.S. The Panel made this decision for fear that the highly variable estimates may undercut confidence in Panel guidance, adversely affecting societal acceptance of their policy recommendations. This concern was clearly expressed in Panel memos (e.g., Crow letters to Weaver, March 7, 12, and 29; May 21, 1956a-d (Crow, 1956a, 1956b, 1956c,1956d)).

Similarly flawed leadership is also displayed in the Panel's subsequent refusal, again by a vote, to provide a written scientific basis for why it recommended the switch to LNT (G.B. Beadle memo to Panel, August 6, 1956 (Beadle, 1956); G.B. Beadle memo to Dr. Detlev Bronk, September 11, 1957 (Beadle, 1957)). Such arrogance with its lack of scientific, societal and historical documentation responsibilities, and improper manipulation of scientific and public policy issues was permitted by NAS president Bronk. While these examples highlight significant biased actions by the Panel, Beyea fails to acknowledge the conflicted roles of Bronk, the chairing of the Panel by Weaver and the awarding of grants to Panelists by those influencing, administering and/or directing Panel activities (Calabrese, 2015a). Beyea also ignored the uniform selection of geneticists for the Panel with the same dose response belief system. The process also failed to offer an independent peer review, resulting in the reinforcing and ensuring of biased perspectives.

3. Did the Genetics Panel have long term influence?

Beyea argues next that the historical and scientific reach of the Genetics Panel was truncated, displaying the equivalent of a short scientific half-life, barely affecting subsequent NAS panels and the national agenda. He fails to grasp the key concept that “as the twig is bent, so grows the tree”. The most significant long-term influence of the BEAR I Genetics Panel occurred when the U.S. National Commission for Radiation Protection and Measurements (NCRPM) generalized the application of the LNT for reproductive to somatic cell mutation within approximately two years after the BEAR I report (NCRPM, 1960; Whittemore, 1986). It is important to note here that James Crow, a BEAR I Genetics Panel member from the University of Wisconsin, was also a member of this NCRPM committee. This application to somatic cells led the way for other advisory committees such as the first BEIR Committee (1972) to use LNT for radiation induced cancer risk assessment. Adopting the guidance of the BEAR I Genetics Panel, the BEIR I report (1972) extrapolated by a factor greater than 1000 in dose and by factors from 100 million to a billion in dose rate, from the observed effects to those occurring in the general population (NCRPM, 1975). LNTbased cancer risk assessment would continue in each of the subsequent BEIR Committee reports over the next forty years. In addition to affecting the risk assessment policy for carcinogens, the BEAR I report (1956) would also have an extended impact via its incorporation into multiple editions of leading genetics textbooks (e.g., Herskowitz, 1962, 1965). While the BEAR I Genetics Panel did not specifically address cancer risk, it is the widely recognized catalyst that led to LNT’s adoption and use in carcinogen risk assessment.

The adoption of LNT was the goal of the Genetics Panel. It had been on the radiation geneticists’ agenda for more than a decade (Calabrese, 2009, 2013a). The Panel used every conceivable means to ensure its adoption and continuance. The NAS helped the Panel achieve this goal by arranging publication of the Panel's findings in the journal Science, having a massive parallel public relations initiative to promote the BEAR I project findings, and keeping the Genetics Panel intact as it moved on to BEAR II in the fall of 1956 (George Beadle, Chair) and BEAR III (James Crow, Chair) into the 1960s. As for Panel members themselves, they actively participated in Congressional Hearings in 1957 on this topic (e.g., Crow, 1957), published numerous articles relating to the topic and tried to ensure that national and international advisory committees had their (i.e., BEAR Genetics Panel) perspectives represented with key radiation geneticists placed in dominant positions (Calabrese 2009, 2013b). For example, even as his health was deteriorating in the mid-1960s, Muller was trying to ensure that his replacement on major national and international committees would be by leading radiation geneticists who reflected his views (Correspondence, Muller letter to EE Pochin, February 12, 1965 Lilly Library Manuscripts Department, Indiana University).

BEIR I (1972) Committee, Genetics Subcommittee of 10 was comprised of three members from the BEAR I Genetics Panel (i.e.

Crow, Neel and Russell) with James Crow serving in the important role of Chairman for this subcommittee. The 1972 report shows the key foundational role that the BEAR I Genetics Panel played as the BEIR I (1972) Genetics Subcommittee reviewed the activities and recommendations of BEAR I and their relevance some 16 years later. They reported that the BEAR I Genetics Panel mantra,–that mutations were nearly always harmful, the damage cumulative and irreversible and the dose response linear–remained almost fully intact, and serving as the key guidance document/blueprint for BEIR I. The only change was that BEIR I determined that the dose response was significantly affected by the dose rate, thereby contradicting BEAR I Genetics Panel. The findings by 1972 indicated that as the dose rate decreases the mutation rate also

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 1. BEIR dose rate graph 1972. Hypothetical dose-response curves for leukemia and genetic effects. (Source: BEIR I, 1972 – page 98). Solid line ¼ observed. Dashed extension of solid lines ¼unobserved. Lines “a” and “b”: possible dose-response curves at high doses and dose rates. Parallel dashed lines ¼ rough limits of error for lines a and b.

decreases as compared to the same total dose applied at a higher dose rate (Russell et al., 1958). In the case of mouse oogonia the occurrence of mutations could be reduced to control levels with a sufficiently low dose rate, supporting a threshold model. In the case of males, the decrease in spermatogonia mutations was also marked and significant but did not return to background levels at the dose-rates tested. BEIR I illustrated that the hypothetical risks of leukemia and genetic damage would be markedly lower at the same total dose, but with a lower dose rate (Fig. 1). The risks for the female are shown to follow a threshold dose response model.

The issue of dose rate was the subject of Ernest Caspari's research with Curt Stern in the late 1940s. Caspari also significantly challenged the concept that ionizing radiation induced mutation was not affected by the dose rate. However, Muller, Stern and others prevented acceptance of the novel Caspari findings by deceptive means as described below since it had the potential to challenge significantly the LNT concept (Calabrese, 2013a, 2015a).

The net impact of the new findings (Russell et al., 1958; BEIR, 1972) that supported the earlier marginalized Caspari research (Caspari and Stern, 1948) led to a revised view that the BEAR I Genetics Panel 1956 estimates of genetic risk “were too high” and on the “conservative side” (BEIR, 1972). However, despite such an acknowledgement, the BEIR I Genetics Subcommittee (1972) would not veer from the LNT recommendation of the BEAR I Genetics Panel, again showing its long term impact and the sustaining leadership of Crow. This decision of the BEIR I (1972) was extended to the BEIR III Genetic Subcommittee (1980) based on the same information, again showing the extended influence of the BEAR I Genetics Panel and its leadership.

According to BEIR 1972, radiation standards used by the U.S. Federal Government were based on the recommendations of the Federal Radiation Council (FRC) (FRC, 1960). The FRC developed the Radiation Protection Guide that is defined as “the radiation dose which should not be exceeded without careful consideration of the reasons for doing so; every effort should be made to encourage the maintenance of radiation doses as far below this guide as practicable.

The BEIR 1972 report states that the present Radiation Protection Guide for the general population was based on genetic considerations and conforms to the BEAR Committee recommendations….” (see page 1, BEIR I 1972 report). Again, the record shows that Beyea incorrectly characterized the impact of the BEAR I Genetics Panel.

In its 1954 Recommendations, the ICRP (1955) struggled with the question of the nature of the dose-response in the low dose

3

zone for mutation and cancer. According to Clarke and Valentin (2009), in their history of the ICRP, the Commission did not reject the possibility of a threshold for stochastic effects. However, a decade later the ICRP (1966) adopted the view that the “Commission sees no practical alternative for the purpose of radiological protection, to assuming a linear relationship between dose and effect, and that doses act cumulatively.” Despite this acceptance of the BEAR I Genetics Panel mantra, the Commission was still not fully persuaded that LNT was scientifically sound, noting that the assumption of no threshold and complete additivity of all doses (i.e. the cumulative and irreversible assumptions of the BEAR I Genetics Panel) may not be correct. Nonetheless, they were satisfied that this approach would not lead to an underestimation of risks.

Clarke and Valentin (2009) indicated that the 1966 publication of the ICRP revealed an important turning point in that it had adopted a precautionary philosophy that was “definitely shaped by the assumption of a linear dose-response relationship without any threshold.” In their deliberations they were strongly influenced by Muller’s concept of “genetic death” that pervaded the BEAR I Genetics Panel (1956) report, so much so that they even adopted the same challenge to estimate genetic damage out to 10 generations, just as Chairman Weaver challenged the Genetics Panel members to do a decade earlier. Thus, the influence of the BEAR I Genetics Panel on the ICRP is apparent with respect to language, concepts adopted, assessment methodologies employed and precautionary philosophy guidance followed.

The Genetics Panel was not concerned with which biological model the LNT recommendation was based on (e.g. fruit fly, bacterial, mouse, etc.); their goal was about controlling and directing the future national risk assessment and management agenda. In fact, Muller engaged in contentious debates in the early 1960s following the discovery of DNA repair/dose rate effects, trying to ensure that these fundamental challenges to several of the tenets of the Geneticists’ Creed would not alter adherence to the LNT (i.e., multiple letter exchanges between Muller and Russell, KZ Morgan and others, 1963). True soldier to his cause, Muller showed his long recognized fierce determination not to yield in debate or to compromise. In fact, a final “golden” moment for Muller occurred during this time period when he successfully fought off a major challenge to LNT based on the concept of DNA repair as lead by Russell et al. (1958) at Oak Ridge, Plough (1962) at Amherst College and others, resulting in the LNT baton being successfully passed on and adopted by the first BEIR Genetics Subcommittee in the early 1970s. Thus, for Muller and his radiation geneticist colleagues, the issue was not about the fruit fly, but about winning on a far deeper and more significant plane.

BEAR I provided the radiation genetics community a major victory along with a platform to secure and extend their influence in affecting textbooks, the research agenda, and national risk assessment policy even as the science would progress. Furthermore, the Panel’s reach continued to affect regulatory perspectives in explicit and important ways. For example, in the mid-1970s, more than two decades after BEAR I, the U.S. EPA used recommendations of the Genetics Panel and its impact on the Atomic Energy Commission (AEC) to adopt the LNT for radiation and chemical carcinogen risk assessment as recounted by Albert (1994), who chaired the EPA Carcinogen Assessment Group during this critical period.

The reach of BEAR I, therefore, was neither ephemeral nor weak, affecting the most fundamental policy decision of EPA in its nearly 50-year history.

Some members of the BEAR I Genetics Panel (Crow, Hollaender, Neel, and Russell) would continue to extend their influence, helping to create the Environmental Mutagen Society in 1969. By 1975, the Society had created a committee (i.e., Committee 17), a goal of which was to establish the mutagen equivalent of the Delaney Amendment (Committee 17, 1975; Drake, 1975, 1978;

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

4

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Drake et al., 1975). On the Committee were James Crow, James Neel, and Alexander Hollaender. Committee 17 drew heavily on both the BEIR I (1972) committee for guidance on exposure limits and the BEAR I Genetics Panel's use of the doubling dose concept which itself was based on its LNT assumption and recommendation. Committee 17 extended this concept to include chemical mutagens, as well as, ionizing radiation. While Committee 17 was never successful in achieving the goal of adding an amendment to the Delaney Clause or creating its equivalent, its efforts were broadly influential as seen in its impact on the International Commission for Protection Against Environmental Mutagens and Carcinogens (ICPEMC, 1983). This serious historical episode with a major environmental mutagen society is notable, showing the long and diverse influence of the BEAR I Genetics Panel.

The impact of the BEAR I Genetics Panel was passed along to the next generation of radiation geneticists. For example, a graduate student of Muller and longtime friend and colleague of James Crow at the University of Wisconsin, Abrahamson (1996), indicated that Crow was responsible for his involvement with UNSCEAR which then led to memberships with the NCRP and ICRP and also with the US NAS Safe Drinking Water Committee (SDWC) and others. According to Russell (1977) Abrahamson would use these appointments to impact the risk assessments of major advisory committees such as Nuclear Energy Policy Study Groups (1977) and the NAS SDWC (1977) in a manner which Russell (1977) believed was “not a reliable one for estimating hazards”, leading to a significant overestimation of genetic risks based on how to interpret the mouse oocyte data of Russell which indicated a threshold dose response. While the BEIR (1980) report favored the interpretation of Russell, the debate amongst the principals would continue into the 1990s. Crow would also influence the selection of another University of Wisconsin geneticist (i.e., Carter Denniston) to multiple NAS radiation committees. Crow et al.

(2005) indicated that much of the risk assessment sections of these expert committee reports were written by Denniston. The issue here however is that the Muller, Crow and others of the original Genetics Panel found ways to extend their influence, casing a long shadow.

There is therefore convincing evidence that the three central themes of the Beyea letter are untenable. The BEAR I Genetics Panel is shown here to have been highly biased in favor of establishing LNT as the default for genetic/cancer risk assessment and that the Genetics Panel had a significant impact on subsequent national and international Advisory Committees and governmental regulatory agencies. It is, however, unknown whether the LNT would have been accepted without the falsified BEAR I Genetics Panel report and the concerted effort of key panel members to ensure that it become institutionalized. Lacking a precautionary principle based foundational belief such as LNT, the field would have soon had to incorporate fundamental biological concepts such as DNA repair, adaptive response, apoptosis, bystander effects, hormesis and non-stochastic- relatively high threshold dose-based impacts (e.g. promotional stimulation) on carcinogen dose responses, amongst other developments into dose-response model development, all of which have the capacity to significantly affect carcinogen dose responses in the low dose zone often leading to non-linear dose responses. Even by the time of BEIR I (1972), with the chairmanship of James Crow, the genetics subcommittee was forced to backpedal, acknowledging that the Genetics Panel's risk estimates using the LNT model were too conservative and overstated estimated genetic damage. Similar back-peddling was seen in the NCRPM (1960) and Committee 17 1975 reports, both of which had Crow on their committees and also adopted the LNT model, but now with the rationale that it represents a conservative public health protectionist philosophy but lacking an adequate scientific basis. This is in contrast to the

Geneticists’ Creed of the BEAR I Genetics Panel that proclaimed a type of scientific certitude. These developments suggest that whether LNT was adopted or not, was significantly impacted by who comprised the Committee, as well as the state of the science.

Beyea's statement that LNT would have been adopted even if BEAR I had not recommended its adoption is therefore speculative.

4. Beyea’s historiography: information filtering Beyea has a very forgiving quality to his writing and judgments when the actions of individuals, no matter how incorrect, egregious or unethical, agree with his perspective. For example, it made no difference to Beyea that Muller (1950a, 1950b, 1954) deliberately misrepresented the scientific record to marginalize the study of Caspari and Stern (1948) which supported a threshold dose response via a dose-rate effect. Muller repeatedly claimed that the control group that Caspari (Caspari and Stern, 1948) employed was artificially high, leading to an incorrect conclusion supporting a threshold. Data from the literature and Muller’s own published and unpublished works supported the opposite perspective, a position that Muller strongly endorsed in preserved correspondence with Stern that was recently revealed (Calabrese, 2015a). Yet, no one had the courage to expose the dishonesty of Muller or at least to advise their esteemed colleague/Nobel Laurette that he had “gone too far” in defense of LNT. The biases and, in fact, dishonesties, of Muller began far earlier, as seen in his deceptive Nobel Prize lecture (Calabrese, 2011a, 2012). However, despite such examples, Beyea fails to acknowledge the significance of such actions in affecting the adoption of LNT.

5. Curt Stern: the hand is quicker than the eye Beyea provides mutation rates from Stern and his research team, suggesting how such variable data might now be interpreted. However, he neglected to inform the reader of key elements in the Stern story. That is, Stern, Muller, Caspari and Uphoff had characterized the Uphoff control group data in her first two key studies as being inappropriate, making her data uninterpretable, leading her mentor Curt Stern to write in the discussion of their manuscript that this was due to investigator bias, a sharp rebuke to Uphoff, while also protecting himself. Despite this rather striking written criticism, Stern would find a way to rescue the spate of uninterpretable data, for as we would come to learn, he was not only an expert in genetics but in data camouflaging as well, all in an effort to preserve the LNT single-hit theory (Calabrese, 2011b, 2015a).

The initial manuscripts of Uphoff and Stern were then hidden from the scientific community in the classified archives of the U.S. Atomic Energy Commission (AEC) such that very few people would have known of the expressed weaknesses of the Uphoff studies except the coauthors, and perhaps several others who were part of the Stern research team (e.g., Uphoff and Stern, 1947).

When Stern used these findings to support the LNT in a subsequent Science publication (Uphoff and Stern, 1949), he neglected to indicate that only a year before these same data had been judged, in writing, as uninterpretable and affected by investigator bias. Stern was not forthcoming and transparent, quite the opposite. He slipped this important fact past the Science journal editors and readership only to be discovered six decades later (Calabrese, 2011a). Thus, Stern continued to marginalize the findings of Caspari while restoring those of Uphoff in ways that were hard to detect. This web of intellectual deceit culminated in the metaanalysis technical note in Science, a paper with no methods/materials and lacking all supporting documentation and data. He and

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Uphoff promised to provide the missing information in a future paper–that never happened.

Beyea also ignores how Curt Stern tried to support the LNT theory in the publication of Spencer and Stern (1948). The Spencer study had numerous serious methodology limitations as documented in Calabrese (2011a), factors that undercut the validity of findings in the low dose zone.

Beyea also ignored the attempt by Stern to prevent acceptance of the Caspari data while still pushing its publication. The discussion of Caspari and Stern (1948) paper bizarrely indicates that the scientific community should not accept the findings of Caspari until it was shown why Caspari's findings differed from the acute exposure study of Spencer which supported the LNT. Stern knew, of course, that the two studies differed in many meaningful ways (Calabrese, 2011a), precluding an experimental resolution of this issue. Thus, Stern acted to promote Caspari via publication in the journal Genetics, where he was the editor, yet ensuring that his findings would not be widely cited. Finally, it is significant that Beyea repeatedly failed to address the flagrant actions of Stern as described above to promote LNT, especially given the central role that these findings played in the adoption of LNT.

6. Bentley Glass: keeping the flame of BEAR I alive Beyea tried to make sense out of the strikingly incorrect historical remembrance of Glass (1991) for how the Genetics Panel (Anonymous, 1956) derived its population based transgenerational mutation risk estimates. Why would this little incident be of interest in the LNT debate? The Glass comments were important because they showed how one of the original revolutionary band of radiation geneticists, who turned cancer risk assessment upside down, was still keeping the flame burning and the myth of the Genetics Panel consensus alive nearly four decades later. Glass (1991) wrote that Dr. Weaver inspired the radiation geneticists of his Panel to estimate the population risk of a gonadal dose of ionizing radiation for up to the next ten generations. Glass incorrectly stated that Weaver told the experts to go back to their hotel rooms and work out their estimations independently. When they returned to the Panel meeting the next day, he wrote that all the major uncertainties were gone, and the mutation estimates amongst the experts were now converging.

This scientific convergence was critical as it would demonstrate expert consensus by independent analyses and not negotiation, thereby assuring that their LNT recommendation would have credibility. While this story may have been how Glass remembered it, nothing could have been further from the truth. The record shows that the Panel did not go back to their hotel rooms but back home, and some 30 or so days later provided estimates that profoundly differed, denying the possibility of any scientific convergence, threatening to fracture hope that their recommendations would be adopted. Glass sought to sustain the myth that scientific consensus had emerged and that LNT had become the default because independent experts had all converged on the same answer–suggesting that they must all be correct. This is the illusionary belief that must be sustained because it hides the scientific misconduct of the entire Genetics Panel.

Unable to figure a way out of this myth dilemma, yet feeling the need to do so, Beyea (2016) creates an alternative hypothetical mirage in which the 85 year-old Glass confused the above inspired speech of Weaver with yet another speech by Weaver. The problem was that the “inspiration” speech for the key meeting for the genetic damage estimates was on Feb 6, 1956 in Chicago, whereas the comments of Weaver to which Beyea refers related to a Nov. 1955 meeting in Princeton, New Jersey and had nothing to do with

5

the population-based genetic damage estimate activity. Despite this little vignette, the issue is serious since it highlighted a real crisis within the Genetics Panel in the winter of 1956: how to achieve independent consensus when there was massive uncertainty both within and between the experts. If their profound disagreements were to become known, the efforts of the Panel to change risk assessment policies might have been threatened. How the Panel hid this problem from the public and avoided possible rejection of their dose response and risk assessment Coup d’état is next.

7. Deception at the highest level 7.1. The Calabrese proposal Beyea contends that my assertion that there was a “discrepancy”–(i.e., actually a deliberate misrepresentation) in the reported number of panel members who made genetic damage calculations and those reported as doing so in the Panel's Science journal publication in 1956 – was not accurate. He states that “there was no discrepancy according to the historians who have studied the issue.” Let us consider the facts. In the 1956 Science paper of the BEAR I Genetics Panel (Anonymous, 1956) on page 1163, column 3, it states: “Six of the geneticists of this committee considered the following problem: suppose the whole population of the United States received one dose of 10 R of radiation to the gonads.

What is the estimate of the total number of mutants which would be induced by this radiation dose and passed on to the next total generation of about 200 [100 million - subsequent errata correction] million children?” At the onset it should be stated that Chairman Warren Weaver's challenge (February 6, 1956) to the Panel geneticists included their providing hypothetical genetic damage estimates of the U.S. population-based on biological models with which they had experience–out through 10 generations (after having been exposed to a gonadal dose of ionizing radiation in the original parental generation and separately to each generation).

The geneticists had highly specialized expertise in their own areas (e.g. bacteria, paramecia, insect models, human population genetics etc.) and so it was hoped that the converging of similar estimates by high level, diverse and independent experts using different biological models and approaches would show to the scientific community and the public that Panel recommendations were on solid ground. That is, the convergence of scientific truth from multiple vantage points was the goal of the Weaver led activity. This convergence was made all the more likely by requiring that the dose response was linear (Weaver memo to Panel, February 8, 1956a (Weaver, 1956a)). What could go wrong?

The Genetics Panel Science article (Anonymous, 1956) states that six geneticists took up the challenge and provided genetic damage estimates; yet, I possess the expert estimates of nine of the 12 geneticists on the panel. Three panelists refused to provide estimates because of profound uncertainty, arguing that any estimates would not be meaningful.

I reported that panelist James Crow, who was in charge of organizing the estimates and making summary tables and graphs (Weaver, 1956b), made an unauthorized decision to remove three of the remaining nine expert estimates (Crow letter to Weaver, March 12 and 29, 1956b, c). As it was his job to collate the submitted estimates, he was the first to see how much they varied.

The variance was of great concern, as he expressed in three initial letters to Weaver (see March 7, 12, and 29, 1956 Crow letters to Weaver, Crow 1956a, 1956b, 1956c). I argue that this concern led to the decision by Crow, Weaver, and subsequently the entire Panel, to drop the three reports with the lowest damage estimates.

These estimates were those based on bacteria and human data.

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

6

By doing so it would significantly reduce uncertainty amongst the expert estimates ultimately presented.

Crow wrote that the estimates would now only be based on fruit fly and mouse data, violating the plan to evaluate the degree of expert consensus from the use of diverse models and analytical approaches. Panelist Neel (1956a, 1956b) (Neel letters to Weaver, March 16, 1956 and April 6, 1956) saw through Crow's decision to restrict the models to Drosophila and mice, revealing that such actions by Crow would bias the findings even further by improperly giving a false impression of scientific agreement. Neel (1956a, 1956b) claimed that Crow's action would create an illusionary “scientific agreement” because as they were essentially forced to assume the same number of genes, mutation rates and other parameters there was limited independence of thought and procedure, a statement in striking agreement with Crow's March 7, 1956 letter to Weaver. Thus, Crow knowingly biased the assessment to show a high level of panel expert convergence. However, according to Neel (1956a, 1956b) the plan fell apart when no “guidance” was offered on how to estimate upper and lower uncertainty bounds and the “apparent” agreement amongst the experts evaporated.

This explanation is provided again on page 21 of the May 18, 1956 draft of the Genetics Panel Final Report. For his third explanation to account for a missing expert estimate, Beyea states that the Wright estimate on human responses was excluded because Wright withdrew it. Yet, the only evidence cited by Beyea to support this contention was the May 21, 1956 letter of Crow to Weaver in which there is no mention of Wright or his genetic damage estimate. For each of the three excluded or dropped expert genetic damage estimates the explanations offered by Beyea are shown by the factual record to be incorrect.

Beyea’s inaccuracies are strikingly emphasized by the following additional information. While a February 8, 1956 Weaver memo charged the experts to provide genetic damage estimates out to ten generations, the Science journal Genetics Panel publication clearly indicates that the estimates actually addressed those only for the “next” generation, with no mention of ten-generation estimates (page 1163, column 3, Genetics Panel, Anonymous, 1956).

Since the Science paper did not include estimations out to ten generations, it does not follow that estimates would be excluded which lacked such estimates. The evidence shows that the alternative explanation of Beyea cannot be accepted as it is directly and repeatedly contradicted by the factual record.

7.2. Beyea hypothesis 7.3. Revisiting the Calabrese interpretation Beyea’s explanation for why the Genetics Panel reported that only six experts provided genetic damage estimates rather than the nine who actually did will be shown to be incorrect. His explanation involves two components: (1) expert reports that failed to provide estimates out to ten generations were excluded; (2) and that the genetic damage estimate of Sewall Wright was later withdrawn, thereby yielding six estimates of damage in the Science paper. While there were two reports that did not provide genetic damage out to ten generations, there is no evidence that this was the reason that they were excluded. The most conspicuous error of Beyea is that one of these two expert damage estimates was not in fact excluded. This was the case of William L. Russell who did not provide an estimate of genetic damage out to ten generations, but did instruct how this could readily be done in his report. The submitted report of Russell (February 24, 1956) was subsequently adjusted by James Crow to include the ten generation estimate. This Crow-adjusted damage estimate of Russell was included in the first package of expert estimates provided in a March 12, 1956 letter and then again in a May 21, 1956 letter– about two weeks prior to the release of the Final Report of the Genetics Panel–both from Crow to Weaver.

The second “excluded” genetic damage estimate was the Demerec report (Demerec, 1956) based on a bacterial assessment method. In his March 12, 1956 letter to Weaver, Crow states “I haven’t included Dr. Demerec’s estimate on the graph for it, too, is based on quite different assumptions that led a greatly different value than the others obtained.” Specifically, the Demerec genetic damage estimation was much lower than the other eight submitted reports. There is no evidence that the exclusion of the Demerec estimate was related to its failure to include estimates out to ten generations. Thus, Beyea’s explanations to account for the apparent discrepancy of six vs nine expert genetic damage estimates are shown to be in error: the Russell report was not excluded, and in the case of the Demerec report, its estimate was dropped but not for the reason given by Beyea.

Further contradicting the ten-generation exclusion explanation of Beyea is that the estimates of Wright (1956) and Kaufmann (1956), which were excluded, did include ten-generation damage estimates. Crow wrote in a March 29, 1956 letter to Weaver that he planned to include only estimates of Drosophila and mice, thereby excluding those with bacteria and humans. Thus, these exclusions were not due to a lack of damage estimates out to ten generations.

The factual record therefore indicates that while the BEAR I Genetics Panel stated in their Science journal article that only six geneticists “considered the problem”, nine actually did. In fact, the nine geneticists did far more than simply “consider” the problem, each provided detailed written estimates.

The Genetics Panel should have indicated in the Science paper that of the 12 geneticists on the Panel, nine provided estimates in the time allotted. In fact, there are nine estimates of genetic damage for the “next” generation. Why did the Genetics Panel write in the Science article that only six geneticists took up the challenge to provide an estimate of genetic damage in the next generation when nine did?

As noted above, heightened concerns about the genetic damage estimate uncertainty both within and between Panelists occurred soon after Crow received the estimates. In a March 7, 1956 letter to Weaver (Crow, 1956a), Crow wrote “Upon looking at the estimates I realize two things. One is that nobody seems to have very much confidence in them. The second point is that those who arrived at comparable estimates usually did it by comparable procedure so that they are not very independent.” Five days later, Crow wrote to Weaver (March 12, 1956, Crow, 1956b) “The groups differ widely in their confidence in the best estimate, as indicated by their grossly discrepant minimum and maximum estimates.” In his March 29, 1956 letter to Weaver, (Crow, 1956c), Crow stated: “The limits presented on our estimates of genetic damage are so wide that the reader will, I believe, not have any confidence in them at all.” By eliminating the lowest three “best” estimates of damage (for bacteria and human data), Crow reduced the variability. Because he was now using only the six closest best estimates he enhanced the appearance of expert estimate convergence and consensus.

Despite the biased exclusions of Crow, it was still not good enough: too much uncertainty remained. How would the Panel “fix” this next problem? They decided to continue to mislead the scientific community and public further by misrepresenting the degree of variation in the remaining six estimates. In a substantial misrepresentation of the record the Panel reported in the Science paper that the uncertainty range of the remaining six estimates was 100-fold; yet it was actually 750 fold. This range of variation was similar for estimates at the first and tenth generations (Calabrese, 2013a). The record thus reveals the Beyea interpretation to be incorrect.

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

The available information supports the conclusion that the Genetics Panel deliberately misrepresented the scientific record, both with respect to the number of geneticists providing estimates and in the expressed variation in the damage of the six estimates on which the report was ultimately based. The decision to report in the journal Science that only six geneticists provided genetic damage estimates for the next generation–rather than the actual nine–and then to misrepresent the uncertainty of those remaining six are grounds to support an investigation into the research record by the entire Panel for fabrication and falsification. These actions were linked in time, supported in memos and consistent with concerns that the public would not accept their recommendations for policy changes if they knew of the uncertainties amongst the experts. Thus, the Panel had the motive, the means and opportunity to carry out these deceptions. Further supportive of this view is that the Panelists voted not to share their estimates with the scientific community (Weaver to Panel memo, June 6, 1956e (Weaver, 1956e)) and the public, thereby preventing uncertainties and deceits from being discovered. A second vote, this time with the newly reconstituted BEAR II Genetics Panel, rejected a request to provide the scientific basis of their June 1956 Genetics Panel final report (see Calabrese, 2015a for detailed documentation; Beadle letter to Bronk, September 11, 1957).

8. Muller: more problems, more errors Beyea rejects my criticism of Muller concerning dose rates used in the Stern studies (with Caspari, Uphoff and Spencer). I indicated that the lowest radiation dose rate in these Stern studies was 0.00165 R/min while Muller indicated it was o 0.001 R/min. This discrepancy is important to resolve. The dose rate of o0.001 R/ min would substantially extend the fold range in the low dose zone over which ionizing radiation mutational effects would be linear (i.e., exceeding 400,000 fold) (Muller, 1950b). His use of o 0.001 instead of 0.00165 R/min to exaggerate the fold range was a deliberate deception. I stated that this was a purposeful deception

by Muller because he had previously used the value of 0.00165 R/ min (Muller, 1950a). The switch to o0.001 R/min could not have been a mistake because it involved too many typographical changes. In addition, Muller was a close advisor on the project and had carefully reviewed the manuscripts prior to publication. Beyea stated that one could get a o0.001 R/min dose rate by simply dividing the dose rate for the 50 R dose in Caspari and Stern (1948) and Uphoff and Stern's (1949) chronic studies by two; this would yield a dose rate for the 25 R dose of o 0.001 R/min as used in another study (i.e., Spencer). Yet, Beyea incorrectly assumed that the experiment with the 25 R was a chronic study; in fact, it was an acute study (Spencer and Stern, 1948) and their methods section indicates that the dose rate was 15–22 R/min. In the chronic studies of Uphoff and Caspari, the total dose was 50 R, with a dose rate of 0.00165 R/min. By bringing attention to this issue Beyea highlighted not only the deliberate deceptions of Muller to enhance the significance of the LNT model but also his own failure to assess the validity of his own argument which was shown to be false.

9. Was James Crow an expert in risk assessment?

Beyea also tries to rehabilitate the efforts of James Crow to estimate genetic damage which seemed strikingly bizarre and antithetical to long standing risk assessment practices, even approaches used in the 1950s. To briefly summarize, Crow chose three different approaches to provide the “best” (median) estimate of genetic damage and the upper and lower bounds: he used fruit

7

fly data for a lower bound; human data from the Japanese bomb data for the upper bound; and mouse data for the best estimate.

Each of these biological models (fruit fly, mouse and human) might be useful for comparison with models having a best estimate and lower and upper bounds. However, to integrate them together in such a manner is inappropriate, lacking biological plausibility. Beyea's ad hoc justification of Crow's patch-work did not alter the conclusion that Crow's methods were not equipped to deal with this question. Significantly, Beyea also failed to mention that Crow not only inappropriately changed the rules of the genetic damage evaluation processes (as noted above), he violated his own self-imposed decision to use only fruit fly and mouse data when he employed human data for his upper bound value.

10. BEAR I must be correct, the UK Panel agreed with it Beyea takes comfort that a parallel radiation assessment for the UK by the British Medical Research Council (BMRC) was in general agreement with conclusions offered by the BEAR I Committee Genetics Panel. What he fails to note is that leaders of their respective national advisory committees monitored and shared developments of the two expert panels, managing these “independent” approaches. In fact, the principal leaders of the two groups [e.g., Detlev Bronk, USA; Sir Harold Hemsworth, Secretary of the British Medical Research Council and several others (e.g.

Weaver)] held a high level confidential conference on April 5, 1956, helping to anticipate and identify possible significant differences to be addressed or possibly smoothed out, ensuring avoidance of political issues (separate Letters from Weaver to Panel and to Hemsworth, April 6, 1956c (Weaver, 1956c)). A review

by Hamblin (2007) details the extent of coordinated oversight that occurred between the two expert groups. Regardless of the type and extent of involvement of these different committees, the biases, scientific misconduct and self-interested activities of the BEAR I Genetics Panel remain. Furthermore, since the Genetics Panel failed to provide a written scientific justification for their LNT recommendation, it is not possible to understand the biological basis for their decisions, let alone directly compare it with other reports. The comments on this issue by Beyea therefore are not appropriate.

11. It starts at the top While I have criticized Dr. Bronk earlier for the introduction of significant structural biases in the creation of the Genetics Panel, criticism of Dr. Bronk in this section centers on his failure to require the Genetics Panel to provide a written scientific justification for the adoption of the LNT. Even when challenged to do so by a group of biologists after the release of the NAS report in June, 1956, Bronk let stand a vote of the Panel to ignore this request (Beadle to Bronk, September 11, 1957). This failure represented a major loss for the scientific community, the public and the history of science.

Bronk failed to ensure that the Panel was accountable to the scientific community. Again, it is hard to understand how Beyea ignores such leadership failings.

One can see the influence of Muller even in the case of a decision not to publish the scientific documentation supporting the LNT decision. In an August 27, 1956 letter to George Beadle, the new chairman of the BEAR II Genetics Panel, Muller states “it would involve us in a lot of thankless work and disputatious rehash of points we have already considered, as well as in airing our dirty linen before the public unnecessarily. After all, only geneticists would be competent to judge the validity of our technical report and geneticists do not need it because they have access to the technical reports already in

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

8

the literature, some of which are in summary form. So far as I can see, it would be a matter of quarreling over what would be the most important points to put in and to what extent they were valid, things on which I thought we had agreed to disagree. Why, for instance, should I enter into a public dispute with Demerec on whether a bacterial generation should be taken as corresponding just as closely to a human generation as a Drosophila generation does? This is only one little example out of many. If you heard Jim Neel’s paper before the WHO meeting…you will see how radically he and I differ regarding not merely the significance of the Drosophila data for man but even regarding what the Drosophila data show with respect to itself. Similarly, I think I would have to disagree with Wright concerning the frequency and importance of small detrimental mutations as contrasted with the conspicuous ones known as lethals and visibles.”

Muller's statement not only shows his geneticist arrogance, but also his failure to understand the duty of the Panel to the scientific community and the public to show the scientific basis of the LNT decision. To say that geneticists could figure out what the Panel did because the papers (which specific papers?) are in the literature is also ridiculous. These comments of Muller not only serve to highlight the irresponsible decision of the Panel and Bronk, but also suggests the influence of Muller in the matter. As seen in his statements above, Muller was critical of genetic damage estimates of Demerec (with bacteria) and Wright (with humans), and, in a decision that very closely matches the written comments of Muller to Beadle, Crow eliminated these models showing low estimates of genetic damage.

12. Stacking the Deck Was the Genetics Panel stacked? Beyea does not think so. What is a stacked Panel? The deck was stacked in this case because only those geneticists in agreement with the Geneticists’ Creed were appointed. Their beliefs were known prior to being invited to join the Panel and reaffirmed by the Geneticists’ Creed statement of Sonneborn discussed earlier. Such a prior belief strongly influenced the final report and guided the next 60 years of cancer risk assessment. The evidence is clear that the deck was stacked in favor of LNT.

13. Hormesis and BEIR VII Beyea introduces the concept of hormesis into the discussion of the history of LNT and my allegations that the U.S. NAS BEAR I Committee, Genetics Panel committed scientific misconduct in their report in the journal Science. The concept of hormesis was not part of this historical evaluation. However, Beyea felt the need to state that I made a presentation to the BEIR VII committee and that the Committee did not apparently consider my presentation sufficiently persuasive. For the record, the BEIR VII committee was kind enough to invite me to speak and paid for my transportation to Washington, DC. They were interested in learning more about the concept of hormesis, the data supporting it and to ask me questions about it. This presentation of about 75 min, was made in September 2000 (Calabrese, 2000), and the final report of BEIR VII was issued in 2006 (NRC, 2006), some six years later. At the time of my presentation the number of citations in the Web of Science database on a yearly basis for the terms hormesis/hormetic was about 400, while by 2015 it had increased to greater than 7200.

Over the past 15 years much has changed, and the principal criticisms of hormesis, such as those related to mechanism, have been substantially clarified and answered (Calabrese, 2013c). In fact, while the BEIR VII committee did not think that the hormesis

concept was quite ready for prime time in 2000 (and it may not have been), the Nuclear Regulatory Commission (NRC) did think it had achieved this status in 2015 when it accepted the credibility of three petitions challenging the LNT model as the default for radiation risk assessment and recommending the use of hormesis (Calabrese, 2015d). While the scientific community now awaits a decision from the NRC on this matter, the point is that the credibility of the hormesis hypothesis has been markedly strengthened over the past 15 years. While this story is not particularly germane to the debate over the LNT and the 1956 Genetics Panel, it is demonstrative of selective use of evidence by Beyea intended to satisfy a narrative.

14. The genetics panel and self interest Beyea plays down the fact that correspondence between BEAR II Genetics Panelists Demerec and Dobzhansky discussed how members of their research community, including themselves, may intentionally overstate potential harm to frighten the public in order to enhance opportunities to acquire grant support (Calabrese, 2014b). It is stated in these letters that “stretching the truth” would be acceptable if it helped to ensure such awards. Thus, we find written correspondence of members of this elite group of radiation geneticists, and members of the BEAR establishment, stating that they would be comfortable distorting the truth to enhance their personal agenda. Furthermore, the BEAR II Committee Genetics Panel voted to refuse to provide the scientific community and the public with a written scientific justification for why BEAR I recommended the switch to LNT for multiple reasons, including their desire to develop a fundable research agenda as a deliverable product (Beadle letter to Panel, August 6, 1956; Beadle letter to Bronk, September 11, 1957).

14.1. Why Did Dr. Warren Weaver resign as Chairman of the BEAR I Genetics Panel?

Beyea (2016) asserts that “Calabrese also charges, without any evidence or citation, that Weaver resigned from the Committee (Panel) after it issued its report to be able, (to avoid) an obvious conflict of interest, to give grants from the Rockefeller Foundation, of which he was a division head.” In support of my statement is a letter of resignation from the Genetics Panel from Weaver to Bronk on June 13, 1956f (Weaver, 1956f) citing four reasons: 1) it required too much time, 2) he was not qualified, since he was not a geneticist, 3) his relationship might be strained, having served as a referee for intense disputes, especially those involving Muller (although Weaver does not specifically mention his name); and, finally, 4) the possibility of conflicts of interest relating to the Rockefeller Foundation's proposed funding of research by some Panel members. On August 6, 1956 George W. Beadle wrote to the Members of the Academy Genetics Committee as follows: “Bronk has asked me to act as Chairman of the Committee. Warren Weaver points out that one of the questions we have to consider is what research needs doing and that conceivably Rockefeller Foundation dollars might be sought to help do it. Therefore, he feels he should not continue as Chairman or even as a member of the Committee.” These letters provide documentation to support my statement.

15. A history of science perspective on the actions of the Genetics Panel The Genetics Panel's misrepresentation of the scientific record in their estimation of trans-generational genetic risk estimates was discussed in some detail by the historian of science John

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Beatty (Beatty, 2006; Beatty and Moore, 2010). He captured the deliberate misrepresentations with the descriptive title: Masking disagreement among Experts. Consistent with the perspective offered in Calabrese (2015a) and in the present paper, Beatty and Moore (2010) stated that the Genetics Panel transcripts showed that the Panel was “laboring under the assumption that their authority depended on being able to reach consensus……Signs of continued disagreement would further erode the public's trust……

Masking their substantive disagreement…was intended not only to protect their authority, but also to foil the leadership of the Atomic Energy Commission, which liked to stress the disagreement among geneticists…in order to thwart the adoption of new, restrictive maximum dosage limits.” Beatty and Moore (2010) further stated “What is clear is that the geneticists assumed—and were correct in assuming—that their authority would be questioned if they revealed the extent of their disagreements.”

This attitude would come into conflict with the requirement for honest reporting of research findings. The February 5, 1956 Genetics Panel transcripts reveals Chairman Weaver trying to convince the Panel to make a distinction between the geneticists’ traditional concept of “scientific accuracy” and what he called “communicative accuracy.” That is, Weaver realized at the outset of the Panel activities that there could be considerable uncertainty amongst the experts on the risk of radiation-induced damage, making a consensus difficult to achieve. He also knew that sharing large uncertainties with the scientific community would reduce the authority of radiation geneticists on such risk matters while hindering public acceptance of Panel recommendations.

The suggestion of Weaver would prove to be a risky proposition as it created a slippery slope that placed scientists in the position of trying to achieve “policy” objectives at the expense of accurately reporting scientific information. What the BEAR I Genetics Panel showed was its scientists were willing to participate in this procedure and to permit the distortion of the research record in multiple ways to satisfy personal and public policy goals. They also permitted themselves to be manipulated by Weaver, who, acutely aware of the importance of the Rockefeller Foundation funding to radiation geneticists, displayed no reluctance in linking the Panel's success to opportunities of funding for its members, indicating that he would “try to get a very substantial amount of free support for genetics if at the end of this thing we have a real case for it. I am not talking about a few thousand dollars, gentlemen. I am talking about a substantial amount of flexible and free support to geneticists”, (National Academy of Sciences (NAS)/National Research Council (NRC), 1956) – NAS transcripts, February 5, page 35.

As part of his interaction with the Genetics Panel, he prefaced his funding remarks with the statement that “There may be some very practical results – and here is the dangerous remark – don’t misunderstand me. We are just all conspirators here together.”

Weaver was direct and clear in linking the project outcome to the funding interests of the geneticists on the Panel. The coupling of funds and outcome was highly manipulative and inappropriate.

16. Beyea claims that challenges to the LNT were made by physicians and physicists, rarely by geneticists The radiation genetics literature for mutations prior to BEAR I provides ample examples that do not support a linear dose response (see Table 1, page 207) (Calabrese, 2009). Second, several additional studies reported at what dose greater than background radiation, induced mutations might be first observed (Sparrow and Singleton, 1953; Giles, 1940). In general, mutations were not observed even when exposures exceeded 1800-fold greater than background. These findings contradicted the LNT model. Third, at the time of the BEAR I committee the New York Times quoted the

9

work of several radiation geneticists whose findings did not support LNT. Ralph Singleton, a highly regarded plant geneticist, stated in an April 17, 1955 New York Times article (Anonymous, 1955) that “there probably is a safe level of radiation, below which no genetic changes occur. Extensive experimentation with corn have shown that it takes a sizable amount of radiation to produce mutations”.

The NY Times (Hillaby, August 2, 1956) also reported that Dr. William J. Schull found no adverse genetic effects on the offspring of many thousands of Japanese victims of the atomic bomb.

Schull's finding challenged dire predictions of birth defects and other adverse effects from Linus Pauling, Hermann Muller, James Crow and others. The findings of Schull and colleagues were presented at the First International Congress of Human Genetics in Copenhagen, August 1–6, 1956, and were published and cited in subsequent reports of leading advisory groups.

Further, the former colleague of Dr. Muller at Amherst College and fellow radiation genetics graduate student at Columbia, Plough (1962), indicated that low doses of ionizing radiation present far less concern than previously believed. This is due to the discovery that cells can repair damage, including mutations via DNA repair processes. This suggested the existence of a threshold.

Finally, the internationally famous geneticist, Mary Lyon, and several colleagues strongly challenged the capacity of radiation geneticists to make accurate predictions of genetic risks in the low dose zone. These comments were made at the very time the BEAR I Genetics Panel was active. Their quotes will be highlighted in the conclusion of this paper (Carter et al., 1956). Thus, the statement of Beyea is incorrect and broadly misrepresents the field.

17. Beyea states that “NAS panels have never doubted linearity of radiation Induced genetic mutations…”

The BEIR I 1972 Committee stated that the BEAR I Genetics Panel (1956) was too conservative in the estimation of ionizing radiation induced genetic risk. This was due to new research findings that the same total dose delivered with a lower dose rate induced less damage than when given more acutely. Taken literally, this would mean that BEAR I Genetics Panel “overstated the genetic risk”. This created the need for a dose-rate modifier to adjust downward risks from a linear model based on acute exposure data as was seen with both the Environmental Mutagen Society Committee 17 (1975) and in the BEIR III report (1980).

The 1980 BEIR III panel was highly conflicted on the nature of the dose response and much has been written about this e.g.

Fabrikant (1980), Radford (1980), and Rossi (1980). It is fair to say that the 1980 Committee was split on the issue of LNT versus threshold.

The 1982 NAS Safe Drinking Water Committee expressed skepticism about LNT modeling noting “…more confidence could be placed in mathematical models for extrapolation if they incorporated biological characteristics of the animal studies…since the users of this volume will be likely to favor different varieties of the conventional extrapolation models or will have access to some of the newer developmental methodologies, it is premature at this state to recommend any single approach by selecting it or calculations…” (page 8, NAS SDWC, 1982).

Model uncertainty of the dose-response in the low-dose region (defined as uncertainty relating to key elements of biological plausibility) has been gaining popularity in NAS publications on risk assessment since the publication of the Red Book in 1983 (NRC, 1983). For example, see the following quote from the U.S. NAS 1994 Science and Judgment in Risk Assessment (NRC, 1994) (where the words “model uncertainty” first appeared) as an example: “The dose-response step involves considerable uncertainty, because the shape of the dose-response curve at low doses is not

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

10

derived from empirical observation, but must be inferred from theories that predict the shape of the curve at the low doses anticipated for human exposure. The adoption of linear models is based largely on the science-policy choice that calls for caution in the face of scientific uncertainty. Models that yield lower risks, indeed models incorporating a threshold dose, are plausible for many carcinogens, especially chemicals that do not directly interact with DNA and produce genetic alterations.” These examples demonstrate that the statement of Beyea is historically incorrect.

18. Beyea states that “The reports of the panel were particularly newsworthy and influential in affecting public opinion because they appeared to contradict the position of the US Atomic Energy Commission (USAEC) that there were no health issues associated with radiation fallout and possibly that there would be a hormesis effect (Creager, 2015; Hamblin, 2007; Seltzer, 2007).”

While AEC officials down played potential adverse human health effects from atmospheric fallout leading up to and immediately after the BEAR I report, no support can be found for the statement by Beyea that the AEC suggests that fallout might affect an hormetic response, based on the three references Beyea cited.

19. What are the documents used by Calabrese to support his historical assessment?

My historical assessment involved identifying and obtaining original documentation and then by purchasing such documents from University Libraries and other repositories of historical documentation (e.g., American Philosophical Society). Thousands of relevant correspondence, memos, manuscripts, and unpublished notes were reviewed, as well as copies of original data notebooks of some leading radiation geneticists as well as relevant published journal articles, dissertations, and monographs. Comprehensive information on all members of the BEAR I Genetics Panel and many close research and professional colleagues and other relevant individuals was obtained from research repositories and reviewed. All these references are in the public domain, but are often bound by copyright laws that affect the capacity for a researcher to make publicly available as Beyea suggests. The archives of the U.S. National Academy of Sciences were also made available for this research.

20. Beyea states that my claims are based on an “incomplete assessment of the historical record” Of particular importance is not what Beyea said but what he did not say. That is, he did not say that I based my assessment on inaccurate statements. If he could have he would have done so.

Instead, he uses the phase “incomplete” assessment without providing new information, context or examples. In fact, he makes such a statement while later admitting his own need to obtain copies of the original documents fully cited in my paper, thereby indicating a limitation of what constitutes the historical record. His comments therefore lack the necessary specificity for informed criticism, being designed to create doubt while trying to avoid being criticized for errors on specific points. This is a standard approach that is designed to mask false argumentation.

21. Discussion Beyea's challenges to my historical account of the regulatory

and public health communities’ adoption of the LNT model reveals that his perspective is highly skewed and inconsistent with the documented historical record. In multiple cases his analysis has significant documented factual errors, for example: his reinterpretation of why the Genetics Panel reported that six rather than nine geneticists provided population genetics damage estimates; his statements regarding Muller's change of dose rate figures for mutational data of Stern; his attempted reconstruction of the Bentley Glass statements and others. In many cases where his technical analysis is not overtly incorrect, he deliberately chooses to tell only part of the story or fails to adequately research the question, filtering facts to fit a biased narrative. This was shown in the case of Curt Stern and the Caspari/Uphoff papers, in the defense of Dr. Bronk (who would not require the Genetics Panel to publish its scientific justification for the LNT recommendation); in the discussion of why Weaver resigned as Chair of the Genetics Panel; and the other examples documented in this article. In addition, Beyea demonstrates a lack of objectivity regarding these assessments. For example, despite overwhelming evidence to the contrary, he continues to state that the Genetics Panel was not biased on the issue of the LNT for genetic damage estimates and that the Panel has had little influence on future dose response/risk assessment policy.

Beyea criticizes me for offering biting criticisms of some revered and extremely accomplished leaders of the radiation genetics community as well as the National Academy of Sciences and one of its past presidents. When one considers the significance and purposeful coordination of their deliberate deceptions, my characterizations must be said, unfortunately, to fit the facts. The fields of toxicology, radiation genetics, and risk assessment were strongly influenced by these esteemed leaders who used their achievements, influence, and access to power to transform policies via deceit–or at least, in part, to allow such transformations through abdication of their responsibilities. It is important that their behavior be judged for what it is. With the co-option of the U. S. NAS these activities would eventually extend to the highest institutional levels spreading the deceptions across continents and generations.

The history of the U.S. NAS has been far from ideal in many respects (Parascandola, 2007). On the personal level, I was involved with the issue of groundwater pollution at the Rocky Mountain Arsenal near Denver, Colorado from 1988 to 2002. In this case the U.S. Army was responsible for vast pollution, in part from the making of the nerve gas Sarin during the 1940s and 1950s. An unintended byproduct in this process was a contaminant called DIMP (dimethyl isopropyl methyl phosphonate) that, because it migrates quickly in groundwater, had moved beyond the Arsenal boundaries into the drinking water of some residents. This led the State of Colorado to enter into legal conflict with the Army and Shell Chemical Company over many issues including potential risks from DIMP. While I was hired to consult for the State the U.S. NAS Committee on Toxicology advised the Army. The NAS selected a toxicologist to chair a committee to evaluate this massive toxicological study of DIMP. This NAS appointed chair had published nearly 20 papers with the lead author of the new study he was appointed to evaluate, including four papers in press at the time of the review. I pointed out to the NAS administration that this was improper, reflecting serious potential bias and a conflict of interest, that the NAS review process was flawed, and that their assessment needed to be redone with a new committee. They refused to acknowledge their errors and strongly disagreed with my recommendation. Furthermore, the NAS committee failed to uncover the research team's apparent falsification of the record (having removed key data on mortality and having changed the wording on how many interventions were employed to prevent mortality in the animals under study) (Calabrese, 2005).

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

The NAS would testify at the major public hearing in 2002 on DIMP and contend that their committee did not have a conflict of interest and that the assessment of the committee was not flawed.

In the end, however, the State of Colorado prevailed discrediting the NAS in the process.

Even though I also have had multiple committee assignments in the past with the NAS (e.g. Safe Drinking Water Committees, Air Cabin Safety, etc.) I have also seen first-hand that the NAS could be corrupted by arrogance, as on the DIMP issue, and just as it has been in the case of the BEAR I Genetics Panel. The NAS strategy in such cases is to appeal to its authority while attacking the messenger. This has certainly been true in the cases of DIMP and the BEAR I Genetics Panel (with which I have considerable experience). Beyea's narrative is not tempered by the reality that the NAS–like all human institutions– suffers both institutional and personal weaknesses, and that it can be wrong, arrogant, and intimidating to critics.

11

this belief and deliberately mislead the scientific community to promote their agenda and careers.

Acknowledgements Effort sponsored by the Air Force Office of Scientific Research, Air Force Material Command, USAF, under grant number FA955013-1-0047 (to EJC). The U.S. Government is authorized to reproduce and distribute for governmental purposes notwithstanding any copyright notation thereon. Long-term research activities in the area of dose response have been supported also by awards from ExxonMobil Foundation (115-1289/138966) over a number of years. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsement, either expressed or implied, of the Air Force Office of Scientific Research, U.S. Government, or Exxon Mobil Foundation.

22. Conclusions The letter by Beyea is highly filtered and contains multiple significant factual errors. None of its arguments are sustained upon analysis. In the end, the letter of Beyea is an example of an effort that generates more heat than light, infusing a spate of new errors as well as more bias and excessive speculation into the literature. The evidence strongly supports the conclusion that the U.S. NAS BEAR I Committee Genetics Panel should be evaluated for scientific misconduct as a result of its deliberate misrepresentation of the scientific record to mask expert uncertainties in their genetic risk assessment estimates.

An insight into some of the views of other groups of geneticists at the time of the BEAR I Genetics Panel meetings may be seen in the statement below that was first read to the British Institute of Radiology Annual Congress, November, 25, 1954 and then published in the British Journal of Radiology, February, 1956, at the same time the BEAR Panel was convening.

“A good deal is now known about the magnitude of the somatic hazard (from ionizing radiation), but almost nothing quantitative about the genetic hazard.

This fact is widely recognized among geneticists, who would doubtless, if they were concerned solely with the scientific aspects of the problem, prefer to remain silent on the subject about which they are so ignorant. However, there has been an insistent demand for quantitative pronouncements, with the result that a number of calculations have been published in which attempts have been made to evaluate the genetic hazard to man. Their value is doubtful, but they have demonstrated a surprising measure of agreement among geneticists in three respects; first, admission of ignorance; second, recognition that almost any answer can be obtained from the calculations, according to the assumptions made and the numerical values fed in; and third, pessimism. Most geneticists would agree that it is very undesirable that members of a human population should be exposed to more than 25 r until after the breeding age, though it is doubtful if any could give really sound reasons for this belief.

From all this it becomes apparent there is an urgent need for more factual information and less theoretical speculation.”

While these statements discredit the main thrust of Beyea’s position, the comments of Carter et al. (1956) are in striking agreement with the deliberately hidden view of the Genetics Panel that there was much uncertainty in the radiation genetics community concerning the estimation of genetic risks in the low dose zone from ionizing radiation. The significant perspective is that Carter et al. (1956) had the intellectual and moral courage to express this situation whereas the geneticists of the BEAR I Genetics Panel, not only failed to state what they truly believed, they hide

References Abrahamson, S., 1996. 70 years of radiation genetics: fruit flies, mice and humans.

Lauriston Taylor lecture. Health Phys. 71, 624–633.

Albert, R.E., 1994. Carcinogen risk assessment in the U.S. Environmental Protection Agency. Crit. Rev. Toxicol. 24, 75–85.

Anonymous, 1955. Experts Explode Fall-out Myths. The New York Times, NY, April 17.

Anonymous, 1956. Genetic effects of atomic radiation. Summary report of the committee on biological effects of atomic radiation by the National Academy of Sciences, BEAR I Genetics Panel (W. Weaver, Chair). Science 123, pp. 1157–1164 [Erratum, Science 124, 170].

Beadle, G., 1956. Letter to Members of the Academy Genetics Committee (Panel).

California Technical University, Pasadena, CA, August 6.

Beadle, G., 1957. Letter to Detlev Bronk. California Technical University, Pasadena, CA, September 11.

BEAR I Report, 1956. National Academy of Sciences (NAS)/National Research Council (NRC). The Biological Effects of Atomic Radiation (BEAR): A Report to the Public, Washington DC. Beatty, J., 2006. Masking disagreement among experts. Episteme 3, 52–67.

Beatty, J., Moore, A., 2010. Should we aim for consensus? Episteme 7, 198–214.

BEIR Committee Report, 1972. Effects on populations of exposure to low levels of ionizing radiation. Advisory Committee on the Biological Effects of Ionizing Radiation, Division of Medical Sciences. National Research Council. National Academy of Sciences.

Beyea, J., 2016. Response to “On the origins of the linear no-threshold (LNT) dogma by means of untruths, artful dodges and blind faith”. Environ. Res. . http://dx. doi.org/10.1016/j.envres.2016.01.039 Calabrese, E.J., 2000. Radiation Hormesis. BEIR VII Committee meeting. National Academy of Sciences, Washington, DC. September 20, 2000.

Calabrese, E.J., 2005. An allegation of scientific misconduct in the Bucci et al. article concerning the effects of DIMP on mink. Reprod. Toxicol. 19 (4), 443–446.

Calabrese, E.J., 2009. The road to linearity: why linearity at low doses became the basis for carcinogen risk assessment. Arch. Toxicol. 83, 203–225.

Calabrese, E.J., 2011a. Muller's Nobel lecture on dose-response for ionizing radiation: Ideology or science? Arch. Toxicol. 85, 1495–1498.

Calabrese, E.J., 2011b. Key studies used to support cancer risk assessment questioned. Environ. Mol. Mut. 52, 595–606.

Calabrese, E.J., 2012. Muller's Nobel Prize lecture: when ideology prevailed over science. Toxicol. Sci. 126, 1–4.

Calabrese, E.J., 2013a. Origin of the linearity no threshold (LNT) dose response concept. Arch. Toxicol. 87, 1621–1633.

Calabrese, E.J., 2013b. How the US National Academy of Sciences misled the world community on cancer risk assessment: new findings challenge historical foundations of the linear dose response. Arch. Toxicol. 87, 2063–2081.

Calabrese, E.J., 2013c. Hormetic mechanisms. Crit. Rev. Toxicol. 43, 580–606.

Calabrese, E.J., 2014a. Response to Letter of Ralph J. Cicerone and Kevin Crowley regarding “How the US National Academy of Sciences misled the world community on cancer risk assessment: new findings challenge historical foundations of the linear dose response. Arch. Toxicol. 88, 173–177.

Calabrese, E.J., 2014b. The genetics panel of the NAS BEAR I Committee (1956): epistolary evidence suggests self-interest may have prompted an exaggeration of radiation risks that led to the adoption of the LNT cancer risk assessment model. Arch. Toxicol. 88, 1631–1634.

Calabrese, E.J., 2015a. On the origins of the linear no-threshold (LNT) dogma by means of untruths, artful dodges and faith. Environ. Res. 142, 432–442.

Calabrese, E.J., 2015b. An abuse of risk assessment: how regulatory agencies improperly adopted LNT for cancer risk assessment. Arch. Toxicol. 89, 647–648.

Calabrese, E.J., 2015c. Cancer risk assessment foundation unraveling: new historical

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i

12

E.J. Calabrese / Environmental Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

evidence reveals that the US National Academy of Sciences (US NAS), Biological Effects of Atomic Radiation (BEAR) Committee Genetics Panel falsified the research record to promote acceptance of the LNT. Arch. Toxicol. 89, 649–650.

Calabrese, E.J., 2015d. Model uncertainty via the integration of hormesis and LNT as the default in cancer risk assessment. Dose-Response 13 (4), 1–5.

Carter, T.C., Lyon, M.F., Phillips, R.J.S., 1956. Induction of mutations in mice by chronic gamma irradiation: interim report. Br. J. Radiol. 29, 106–108.

Caspari, E., Stern, C., 1948. The influence of chronic irradiation with gamma-rays at low dosages on the mutation rate in Drosophila melanogaster. Genetics 33, 75–95.

Clarke, R.H., Valentin, J., 2009. The history of ICRP and the evolution of its policies.

Ann. ICRP 39, 75–110.

Committee 17 (JW Drake, Chairman), 1975. Environmental mutagenic hazards.

Mutagenicity screening is now both feasible and necessary for chemicals entering the environment. Science 187, 503–514.

Creager, A.N.H., 2015. Radiation, cancer, and mutation in the atomic age. Hist. Stud.

Nat. Sci. 45, 14–48.

Crow, J.F., 1956a. Letter to Weaver. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, March 7.

Crow, J.F., 1956b. Letter to Weaver. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, March 12.

Crow, J.F., 1956c. Letter to Weaver. Lilly Library (Sonneborn mss). Manuscript Department. Indiana University, Bloomington IN, USA. March, March 29.

Crow, J.F., 1956d. Letter to Weaver. Lilly Library, (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, May 21.

Crow, J.F., Testimony - Statement of Dr. James F. Crow, Professor of Genetics and Zoology. University of Wisconsin. Hearings before the Special Subcommittee on Radiation of the Joint Committee on Atomic Energy. In: Proceedings of the Congress of the United States, 85th Congress, 1st Session, Part 1. 1957 United States Government Printing Office, Washington DC. Crow, J., Engles, W., Susman, M, Temin, R. (Memorial Committee), 2005. Memorial Resolution of the Faculty of the University of Wisconsin-Madison on the Death of Professor emeritus Carter Denniston. November 7, Faculty Document 1871.

University of Wisconsin, Madison.

Demerec, M., 1956. Letter to Weaver. Lilly Library, (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, February 14 with February 11 Document..

Drake, J.W., 1975. Environmental mutagenesis: evolving strategies in the USA. Mut. Res. 33, 65–72.

Drake, J.W., 1978. Some guidelines for determining maximum permissible levels of chemical mutagens. In: Flamm, W.G., Mehlman, M.A. (Eds.), Advances in Modern Toxicology, Volume 5-Mutagenesis. Hemisphere Publishing Corporation, New York, p. 926.

Drake, J.W., Abrahamson, S., Crow, J.F., Hollaender, A., Lederberg, S., Legator, M.S., Neel, J.V., Shaw, M.W., Sutton, E.E., von Borstel, R.C., Zimmering, S., de Serres, F. J., 1975. Environmental mutagenic hazards. Science 187, 503–514.

Fabrikant, J.I., 1980. The BEIR III controversy. Rad. Res. 84, 361–368.

Federal Radiation Council (FRC), 1960. Background material for the development of radiation protection standards. Federal Radiation Council, Washington DC, Report No. 1.

Giles, N., 1940. Spontaneous chromosome aberrations in Tradescantia. Genetics 2, 69–87.

Glass, B., 1991. The Rockefeller Foundation: Warren Weaver and the launching of molecular biology. Quart. Rev. Biol. 66 (3), 303–308.

Hamblin, J.D., 2007. A dispassionate and objective effort: negotiating the first study on the biological effects of atomic radiation. J. Hist. Biol. 40, 147–177.

Herskowitz, I., 1962. Genetics, First Edition Little Brown & Company, Boston MA. Herskowtiz, I., 1965. Genetics, Second Edition Little Brown & Company, Boston MA. Hillaby, J., 1956. Geneticists Find no Atomic Harm. The New York Times, New York, NY, August 2.

International Commission for Protection Against Environmental Mutagens and Carcinogens (ICPEMC), 1983. Committee 4 final report. Estimation of genetic risks and increased incidence of genetic disease due to environmental mutagens. Mutat. Res. 115, 255–291.

Kaufmann, B.P., 1956. Letter to Weaver. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, February 27.

Muller, H.J., 1950a. Some present problems in the genetic effects of radiation. J. Cell.

Comp. Physiol. 35 (Suppl. 2), S9–S70.

Muller, H.J., 1950b. Radiation damage to the genetic material. Am. Sci 38, 32–59, 126.

Muller, H.J., 1954. The nature of the genetic effects produced by radiation. In: Hollaender, A. (Ed.), Radiation Biology 1. McGraw-Hill Book Company, New York, pp. 351–473.

National Academy of Sciences/National Research Council, 1956. BEAR I Committee Genetics Panel Transcript.

National Academy of Sciences Safe Drinking Water Committee (NAS SDWC), 1977, Drinking Water and health. Vol 1, National Academy of Sciences, Washington DC. National Academy of Sciences Safe Drinking Water Committee (NAS SDWC), 1982.

Drinking Water and Health. Vol 2. National Academy of Sciences, Washington DC. National Committee on Radiation Protection and Measurements (NCRPM), 1960.

Report of the Ad Hoc Committee. Somatic radiation dose for the general population. Science 131, 482–486.

National Council on Radiation Protection and Measurements (NCRPM), 1975. Review of the current state of radiation protection philosophy. Recommendations of the NCRPM. NCRP Report No. 43. Bethesda, MD. National Research Council (NRC), 1983. Risk Assessment in the Federal Government: Managing the process. National Academies Press, Washington DC, p. 191.

National Research Council (NRC), 1994. Science and Judgement in Risk asssessment.

National Academies Press, Washington DC, p. 672.

National Research Council (NRC), 2006. BEIR VII: Health Risks from Exposure to Low Levels of Ionizing Radiation. National Academies Press, Washington DC. Neel, J.V., 1956a. Letter to Weaver. American Philosophical Society (APS). University of Wisconsin, Philadelphia PA, USA, March 16.

Neel, J.V., 1956b. Letter to Weaver. American Philosophical Society (APS). University of Wisconsin, Philadelphia PA,, USA, April 6.

Nuclear Energy Policy Study Group, 1977. Nuclear Power Issues and Choices. Ballinger Publishing Co, Cambridge, MA. Parascandola, M., 2007. A turning point for conflicts of interest: The controversy over the National Academy of Sciences’ first conflicts of interest disclosure policy. J. Clin. Oncol. 25, 3774–3779.

Plough, H.H., 1962. The harmful genetic effects of radiation. J. Natl. Med. Assoc. 54 (6), 652–657.

Radford, E.P., 1980. Human health effects of low-doses of ionizing radiation. The BEIR III controversy. Rad. Res. 84, 369–394.

Rossi, H.H., 1980. Comments on the somatic effects section of the BEIR III report.

Rad. Res. 84, 395–406.

Russell, W.L., 1956. Report to Weaver. Estimate of Expressed Damage in the Offspring of Individual Exposed to 10r. Lilly Library, (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, February 24.

Russell, W.L., 1977. Mutation frequencies in female mice and the estimation of genetic hazards of radiation in women. Proc. Natl. Acad. Sci. 74, 3523–3527.

Russell, W.L., Russell, L.B., Kelly, E.M., 1958. Radiation dose rate and mutation frequency. Science 128 (3338), 1546–1550.

Sparrow, Ah, Singleton, W.R., 1953. The use of radiocobalt as a source of gamma rays and some effects of chronic irradiation on growing plants. Am. Nat. 87, 29–48.

Seltzer, M.W., 2007. The technological infrastructure of science. Dissertation. Virginia Polytechnic institute and State University, Blacksburg, VA. Spencer, W.P., Stern, C., 1948. Experiments to test the validity of the linear r-dose/ mutation frequency relation in Drosophila at low dosage. Genetics 33, 43–74.

Uphoff, D.E., Stern, C., 1947. Influence of 24-hour gamma-ray irradiation at low dosage on the mutation rate in Drosophila. MDDC-1492, U.S. Atomic Energy Commission. Hathi Trust Digital Library, pp. 1–6 〈http://www.hathitrust.org〉.

Uphoff, D.E., Stern, C., 1949. The genetic effects of low intensity irradiation. Science 109, 609–610.

Weaver, W., 1956a. Memo to Members of the Genetics Panel. Lilly Library. Manuscript Department. Indiana University, Bloomington IN, USA, February 8.

Weaver, W., 1956b. Memo to Genetics Panel, Brief Report. Lilly Library (Muller mss).

Manuscript Department. Indiana University, Bloomington IN, March 2.

Weaver, W., 1956c. Letters to Panel and Hemsworth. Lilly Library. Manuscript Department. Indiana University, Bloomington IN, April 6.

Weaver, W., 1956d. Memo to Panel. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, June 5.

Weaver, W., 1956e. Memo to Panel. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, June 6.

Weaver, W., 1956f. Letter to Bronk. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, June 13.

Whittemore Jr., G.F., 1986. The National Committee on Radiation Protection, 1928– 1960: From Professional Guidelines to Government Regulation. Department of History of Science. Harvard University, Cambridge.

Wright, S., 1956. Letter to Weaver. Lilly Library (Muller mss). Manuscript Department. Indiana University, Bloomington IN, USA, February 22.

Please cite this article as: Calabrese, E.J., LNTgate: How scientific misconduct by the U.S. NAS led to governments adopting LNT for cancer risk assessment. Environ. Res. (2016), http://dx.doi.org/10.1016/j.envres.2016.03.040i