When assessing epidemiology-based health effects of radiation, a critical question is, How good is the dosimetry? The problems in dosimetry are markedly different for Hiroshima and Nagasaki by virtue of two different types of bombs and their different radiation energy releases. Basically one needs to determine the transmission factors in air, the shielding characteristics of the exposed individual, and finally the individual's organ doses, and to make these determinations for over 100,000 people at varying distances and under a variety of conditions.

In the first of two chapters on dosimetry, Kerr traces the evolution of the dosimetry analysis, beginning with studies at the Nevada Test Site undertaken to empirically estimate radiation fields. Subsequent testing of fabricated Japanese houses provided information on transmission factors both inside and outside the houses. Incorporating survivor interview data on location and shielding led to a system of dose estimation known as T65D which was used to estimate and develop health risk analyses for approximately 15 years from its inception. Conflicts between these estimates and newer information led to an international review and the call for a more accurate dosimetry analysis. This reanalysis leading to the incorporation of the Dosimetry System 86 (DS86) is detailed both in the chapter by Kerr and the succeeding chapter by Kaul. In this later chapter an extensive discussion is provided on the new technologies introduced to assess dose; the new dosimetry, primarily by increasing gamma dose in Hiroshima and simultaneously reducing neutron dose contribution led to a total reduced biologically effective dose.

However, both Kerr and Kaul point out that there is still a discrepancy in Hiroshima neutron data, which remains to be resolved, and the extent of the increase in amount of neutrons' impact on dosimetry is unclear. There are additional concerns involving the shielding characteristics of factory workers in Nagasaki since their higher assigned doses do not correlate well with their lower cancer risks. Chromosome analysis of survivors, to be discussed below, also provides additional evidence of discordance between dosimetry and biological effects.

Beginning in the 1950s, it became apparent that the A-bomb survivors would be incurring continuing radiation risks from medical exposures, diagnostic as well as therapeutic procedures, particularly since routine examinations were not only part of the ABCC Adult Health Study (AHS) practice, but community medical examinations were also occurring. Kato's paper deals with radiological surveys conducted in both cities at the ABCC, and local hospitals and clinics beginning in 1961, which attempt to assess the cumulative doses received by the participants in the AHS study. When assessments of the risks of radiation from the atomic bombs are extrapolated into the low-dose range it clearly becomes important to know how much additional dose was received from routine examinations. Initially the number of radiographic examinations steadily increased, but there was a leveling off in the two cities because of the awareness of physicians of the necessity to avoid unnecessary radiation exposure. Nevertheless, the data accumulated both from survey of equipment and interviews have indicated that an appreciable proportion of the AHS participants did receive bone marrow doses in the range of 1–9 cGy by the end of 1952, doses equivalent to that received from the atomic bomb exposures. Just how these doses will be factored into the risk estimates has yet to be determined. The presentation is replete with figures and tables documenting the analyses. Additionally, new data on radiation therapy and second cancers are in the process of analysis.

The Nakamura paper deals with a number of different aspects of cytogenetic analysis of the A-bomb survivors. For the purpose of this preface we limit ourselves to those studies which relate directly to biodosimetry issues. Conventional cytogenetic analysis for stable reciprocal translocations has been completed on more than 2,500 survivors in both cities. There has been a consistent city difference: the Hiroshima dose-response curve is higher and more linear than the Nagasaki curve even with the new DS86 dosimetry. A discrepancy in dose response appears to exist in Nagasaki for those exposed in tenement houses or not in single family dwellings versus those in single family dwellings. No such serious difference exists in Hiroshima; moreover, there is good agreement for data from single family homes between the cities. In addition, chromosome analysis agrees with other biological endpoints in demonstrating that there are large random dose errors when comparisons of supposedly equivalent high-dose individuals with and without severe epilation are studied. In all cases the epilators show a two to three times greater response. These discrepancies are resolved with results from electron spin resonance (ESR) studies on tooth enamel. The ESR studies show high correlation with chromosome aberration data, and both have lower correlations with assigned doses. Therefore it appears that the differences are the result of distance-biased dose estimation, in this case presumably due to improper localization or shielding conditions, and not to differential radiosensitivities.

The next sequence of papers deals with RERF cancer studies, and we choose to start with Pierce's presentation on statistical aspects of this epidemiology program because he has provided an extremely clear historic exposition of the statistical

program, its evolution to fit the needs of the emerging data, and its implications, topped off by examples from the then-unpublished Life Span Study Report 12 on cancer mortality. Statistical analysis has been fundamental to almost all of the ABCC/RERF research programs. Pierce describes the early use of the "contingency table method," wherein the specific exposure categories defined by city, sex, and age at exposure were used to determine both the expected and observed cancers seen, demonstrating that a dose effect was indeed observable. The limitations of this method then led to the next stage, regression models. Again the limitations are clearly presented, particularly the inability to study temporal patterns of excess risk as the follow-up period increased. This in turn led to the excess relative risk model (ERR), which incorporated stratification variables used in the earlier methods. The simplicity and elegance of the ERR model were important attributes for epidemiology. As Pierce points out, caution must be exercised both in asking the right questions and in preventing overinterpretation of the data. He next examines the limitations of this method and moves on to describing the usefulness of excess absolute risk (EAR) methodology, which is playing an increasingly important role in understanding the nature of radiation-induced cancer. He demonstrates the use of the ERR and EAR methods with respect to the recent mortality analysis, where age at exposure and sex suggest apparently greater difference by the relative risk method than the EAR method, in which the small number of background cases in both sexes and in the young may be distorting the results by the ERR method.

Leukemia was the first cancer recognized to be radiation-induced. Preston's paper on leukemia risks in A-bomb survivors traces the history of the leukemia studies, originating with the concerns of two Japanese physicians in Nagasaki, Takuso Yamawaki and Masanobu Tomonaga, in the late 1940s that an excess of leukemia was apparent in the survivors. This led to a number of surveys by the ABCC in 1951, 1953, and 1957. However, until the 1960s there was no consistent standard of ascertainment or review nor a defined population to study. Stuart Finch was responsible for the establishment of the ABCC Leukemia Registry and detailed shielding histories for all survivors within 2,000 meters of the bomb. By the mid-1980s, cases were reclassified by the French-American-British (FAB) system, a detailed classification of leukemia subtypes. The recent emergence of the two city tumor registries now serves as the primary source for new cases. Much of Preston's paper deals with the serial analyses of risks as data accumulated. Significantly, the first quantitative risk estimate was developed in 1957 by E.B. Lewis, who was named a Nobel Laureate during this symposium (for his studies on developmental genetics). The most recent analysis covering the period 1950–1987 is reviewed in the last section. It should also be noted that 50% or more of the cases of leukemia (ALL, AML, CML) are attributable to radiation exposure. Detailed models of all leukemias combined, and for ALL, AML, CML, ATL and other leukemias, are presented, as well as tabular data on changes of excess risk with time. The dose-response curve for combined leukemias is significantly non-linear, i.e., linear-quadratic in shape, with no convincing evidence for a threshold. The data indicate

that while risks have fallen with time, they have continued to persist for those who were exposed as adults. Continued follow-up is required to see whether those who were young at the time of exposure and are now entering the "cancer prone" ages will show additional response.

The establishment of acceptable, high-standard tumor registries in the A-bomb cities was a hard-fought process. Although off to a promising start in the early 1950s, disagreements between important hospitals and members of the medical community and ABCC led to its decline. With the transition to RERF, old hostilities faded and the efforts of a dedicated group of staff reestablished the registries as modern population-based facilities. Mabuchi, a principal architect of this rejuvenation, describes the history, role, and usefulness of this program. The cancer incidence data collection, procedures, and staff were revitalized, and records from 1958 to 1987 were of such high quality that it was possible to carry out a comprehensive analysis. To develop quality data, hospital records must be collected, abstracted, linked to tissue registries, and RERF's Master File of Life Span Study participants, cross-checked with death certificate data and other clinical program data at RERF. Quality is measured by several indices, but high among these is the frequency of histological verification and the mortality-to-incidence rates. The registries now rank among the best in Japan.

While covering a shorter span of years than the mortality studies, the incidence studies yield more cases and more radiation-attributable cancer. Much of this comparative analysis is described in the companion paper by Ron. As should be emphasized, the solid cancer data is remarkably linear over the dose range studied, including very low doses, and the confidence limits of the data are very narrow indeed.

Followed over time, the earlier years were dominated by cancers appearing in the over-20 age-at-exposure group, but almost 50% of all total excess cancers were observed in the last 7 years of follow-up in the under-20 group, a cohort that has 85% of its members still alive and entering the cancer-prone period. The tables presented provide extensive details on sites, dose, excess cases, age, and sex.

There is a useful discussion of the limitations of registry data and the means to overcome such deficiencies.

Examples are given of how the registries are facilitating site-specific, case-control and nested case-control epidemiologic studies. For example, in liver cancer studies, what are the roles of hepatitis B and C viruses in relationship to radiation?

Until very recently, cancer mortality has served as the main indicator of radiation-induced somatic effects, based as it is on death certificate data readily obtainable from all of Japan because of the unique system known as the koseki family registry system. This assures virtually complete coverage. However, death certificate data is recognized to have serious inaccuracies which are circumvented in incidence data collection with its very high histological verification level. A major effort over the last decade has brought the two city tumor registries into a highly respected and accurate state. The paper by Ron et al. summarizes the first major analysis

covering 1958–1987 and contrasts results with the mortality data from 1950 to 1987. It should be pointed out that incidence data is only collected from major hospitals and clinics within the two prefectures and is based on a smaller Life Span Study (LSS) population base. Nevertheless, the number of total cases and radiation-induced excess cases exceeds that in the mortality data by about 23% and 65%, respectively. Because of lower mortality, cancers of the breast, thyroid, skin, and salivary glands contribute to this difference. The two datasets are for the most part comparable with respect to risk of stomach, colon, liver, lung, breast, ovary, and urinary bladder cancer. By either risk analysis, the statistics ERR/Sv or 10⁴ PYSv for the incidence series have larger risks by 40% and nearly threefold higher, respectively. One of the most remarkable features of the data is the exquisite linearity with dose over the wide range studied. Using EAR comparisons, the female risk of cancer in both studies is about twice that of males, and there is a much higher EAR of 22.4 for incidence compared with mortality (3.8) for those exposed under the age of 20. The difference is not as great for those over 20 at the time-of-bombing (ATB). These datasets will continue to play a major role in the evaluation of cancer risks, particularly since over 80% of the under-20 age group are alive and entering the cancer-prone age period over the next two decades. In conjunction with the tumor registry, the tissue registries of the two cities provide important access to the material needed for site-specific studies, which in turn allows more in-depth analysis of cancer-specific subtypes and their radiosensitivity, and facilitates incipient studies on the molecular basis of cancer, and perhaps whether radiation-specific molecular lesions at the DNA level will be recognizable.

Ethel Gilbert's paper on nuclear workers exposed to low-level radiation has immediate relevance to the preceding papers on risk estimates of the A-bomb survivors because it provides a check on the validity of the extrapolations to low dose from the latter study, which have served as a standard for radiation protection regulation over several decades throughout the world. While the presentation examines the individual worker studies of the United States, Canada, and the UK, it is the combined international study carried out by the International Agency for Cancer Research (IARC) that provides the most powerful database on which to compare results for male workers monitored for dose, aged 20 to 60, against the corresponding subset of the LSS data. For example, the estimated ERR/Sv for all cancers excluding leukemia in the international worker study is close to zero, but the upper 95% confidence interval (CI) overlaps that of the A-bomb study; in fact individual national studies can exceed the Japanese study, but there is less precision in such data. For leukemias excluding chronic lymphocytic leukemia (CLL), the two risk estimates were remarkably close: 2.2 for workers and 3.7 for survivors with similar upper 95% CI, but a much lower but non-zero estimate for the lower 95% CI for workers. The remaining discussion deals with uncertainties associated with confounding factors and dosimetry issues which could affect the worker analysis. Given the even more complex issues surrounding the A-bomb studies, it is very

reassuring that the single acute-dose data and much higher overall dose data agree so well with the low-dose and mostly low-dose rate data. The A-bomb data have greater precision at present than the worker data, which is important for a number of reasons, primarily for risk estimation. The combined worker analysis also indicates that the A-bomb data have not underestimated risk at low doses; in fact, the upper confidence limits of the worker study quite effectively rule out underestimates by one order of magnitude, as some investigators have charged. It is clear that continued follow-up of both datasets will make important contributions to the low dose risk assessment issue.

Land's presentation deals with the interaction of radiation dose and other risk factors. He uses smoking/lung cancer and reproductive factors/breast cancer as a two-model system in which to analyze the interactions in the A-bomb survivors. Both cancer endpoints are clearly demonstrated to be linearly dose related. Sex and age have readily been demonstrated to affect some cancer rates and are easily observable risk factors. As is well known, smoking is a principal contributor to lung cancer production in non-radiation-exposed populations; and reproductive performance, age at first birth, number of births, and cumulative lactation history are well-known moderating factors for ''naturally" occurring breast cancer.

For lung cancer, Land points out that heavy cigarette smoking is a much greater risk factor than even the highest dose received by the A-bomb survivors. But the question of interaction deals with how the two agents influence the final response. Will the excess risk be the sum of the two risks, an additive response, or greater than additive, a multiplicative response (a × b), or somewhere in between, as developed by statistical analysis? While the published data at RERF are unable to distinguish between these models, the data of US uranium miners suggest that a mixture model intermediate between additive and multiplicative gives the best fit.

For breast cancer the analyses (based on case-control studies) were much more clear cut. That is, the interaction results consistently conform to the model predicted by the multiplication formula and are significantly different from the additive model. As stated above, age at first pregnancy, number of births, and cumulative lactation time are major interactive factors. Based on rodent experimental studies, there is the hypothesis that cells that may have been initiated toward the cancer states are reprogrammed to differentiated milk producing cells by the pregnancy process, thus removing them from the pool of cells still cancer prone.

In the area of modeling of radiation-induced cancer production, the following series of papers provide different approaches to this issue.

The Mendelsohn paper has developed a modified Armitage-Doll model derived from the RERF solid cancer data. The Armitage-Doll model describes the temporal pattern of many cancers as a power function of age, dependent upon a number of irreversible steps, in the range of five to six in their model. This would now be supported by molecular analysis which has implicated five to six mutational steps in the development of colon cancers. Mutations in oncogenes and tumor-suppressor genes and other types represent the pool from which a subset of mutational events

may result in a cancer. Mendelsohn's model is simply that radiation has a certain dose-dependent probability of introducing a mutational step in the process, thereby reducing the number of mutational steps by one. The model predicts that the number of induced cancers will follow the pattern for the overall non-exposed cancer rate but be accelerated in time by one step. Mendelsohn credits the well-defined linear dose-response observations as critical to the evolution of his thinking. However, it seems to us that it is the induced event (and not the specific dose-response relationship, linear or linear quadratic, for example) that is critical to the model; the simplicity of this central induced event imparts a sense of biological elegance to the model. Caveats are well spelled out in the final section.

As we have seen in the Mendelsohn paper, the Armitage-Doll model provided a mechanistic approach to cancer risk. Little's presentation also utilizes the A-D model as a starting point to his analysis and describes various modifications to the model developed by himself and colleagues as well as other workers. Since the A-bomb study provides the largest source of cancer data, most of the models attempt to fit the mathematical modeling to the observed results in terms of excess relative risk. Little develop his "optimal solid cancer model," a three-stage model, restricted to the first stage being radiation affected (compare to Mendelsohn's unrestricted stage). The leukemia data are also fitted to a three-stage model with the first and second stages responsive to radiation. Extensive discussion of these models is provided, including their limitations. He then reviews the two-stage model of Knudson and the Moolgavkar, Venzon, Knudson (MVK) generalized model and compares the conditions to a generalized multistage model (modified A-D). In the MVK model, the conditions involve the number of stem cells, with varying mutation rates and mutation steps, cell division rates, and elimination rates, while somewhat different elements exist in the multistage model. By varying these parameters, Little attempts to show how they effectively predict cancer observations in the Life Span Study, and he concludes that three or more mutation steps reconcile with the solid cancer epidemiologic data.

We wonder if Occam's razor is not being blunted by so many ways of varying the parameters.

In the paper on the genetic effects of the atomic bombs presented by Neel, the founder of the genetics program at ABCC and first director of the entire program, we are provided with a summary of all the major studies since the inception of the genetics program. Remarkably, Neel and Schull have provided guidance to this, the largest human genetic prospective study of its kind, throughout the life of the program. The paper begins with the early history of data collection on the children of the survivors, some 70,000, of whom 31,150 were born to parents who were exposed. In recent years, reanalysis of the major studies was undertaken with respect to DS86 dosimetry, namely in terms of untoward pregnancy outcomes, mortality of liveborn for the first 26 years of life, tumor development through age 20, cytogenetic abnormalities, and mutations altering the electrophoretic behavior of some 30 different blood proteins. It should be recognized that to date there

has been no significant increase in any of the genetic endpoints studied, consistent with the expectations when the program was established. The doubling dose (DD) concept was used as a measure of radiation damage, i.e., the estimated contribution of spontaneous mutations summed over these endpoints was divided by the sum of the dose-related regression coefficients. This value provides an estimate of risk from acute radiation. The DD ranged between 1.7–2.2 Sv and was assumed to be about twice as large if converted to the chronic exposure condition. Depending on which experimental mouse data are used for comparison, Neel estimates that both humans and mice are much less radiosensitive than previously thought, or that humans are less sensitive than the estimate of 1 Sv DD derived from specific-locus mouse studies. In the present work, which is attempting to assess genetic damage at the DNA level, several techniques are elaborated, including: scoring for nucleotide substitutions using denaturing gradient gel electrophoresis, detection of changes in unique minisatellite DNA sizes, and detection of increases or decreases in thousands of DNA nucleotides. This study may have significance to a new array of genetic diseases involving increases in trinucleotide repeats. Another approach involves two-dimensional DNA gel electrophoresis, which allows optical scanning of thousands of gene fragments and their computer analysis, for parents and children, and which will detect 50% intensity changes in the specific spots. These studies are still in the developmental stages, but await the final resolution of the question, Were mutations induced in the parents' germ cells? Neel also speculates on new DNA approaches to correlate genetic and somatic risk estimates. He ends with two important issues: the approach one would take if these studies were done again and, secondly, the independence of the ABCC studies from any government pressures, because of the buffer provided by the National Academy of Sciences management, an issue not appreciated by revisionist historians.

Trosko's provocative presentation to this symposium is a theoretical examination of the molecular factors involved in the evolution of a cancer. Three main stages are examined: initiation, promotion, and progression before the ultimate appearance of cancer. In order for the final event to occur, the breakdown of a myriad of defense mechanisms must occur at the molecular, cellular, tissue organ, and organ system levels in a creature such as ourselves. One issue that Trosko is primarily concerned with is at what stages low radiation levels (i.e., very low doses) can affect the enormous variety of defense mechanisms that have evolved in aerobic organisms to protect against oxidative stresses from both endogenous metabolic pathways and exogenous agents. While there is no question that radiation can induce mutations (i.e., primarily deletions) in DNA of recessive tumor-suppressor genes and rearrangements of dominant acting oncogenes, Trosko asks whether radiation is capable of affecting where promotion and progression stages occur. He presents an enormous array of cellular events that would indicate low-dose radiation would not act as a tumor promoter. The question is not completely resolved (to our minds) because, among the cascade of events that could result from even a single ionization track in non-genetic cellular regions, the possibility that some

epigenetic events can affect the stem cell of interest and its cellular descendants in ways that can be considered promotional is not resolved, particularly since every cell of interest may be in different metabolic susceptibility states. Trosko raises these points of concern as well. A major discussion in this paper is the important role of intercellular communication in maintaining cellular homeostasis, thereby preventing metabolic disruptions. It appears that many stem cells, however, do not express the gap-junction genes that initiate intercellular communication. Trosko argues that sustained chronic exposure could only bring about promotional activity if a series of conditions regarding oxidative stress could be met that would exclude cell death and allow cell division. The essence of his analysis is that it is unlikely that low doses of radiation can affect all the steps necessary to the development of a malignant cancer. The reader can contrast this view with that presented in the paper by Mendelsohn reviewed earlier; then the question becomes: Is it necessary for radiation to affect all these steps, if it can advance the process by just one step? This brief review does not do justice to Trosko's elaborate presentation, for which we apologize.

It would be presumptuous of us to try to summarize the very difficult concepts presented by Bond, wherein he compares the factors involved in a classical medical pharmaceutical dose response in an individual, of a threshold sigmoid shape, with the population-based epidemiological response to cancer from radiation. The measure of dose on a population basis is rederived. From population to organ-tissue to cell—Bond concludes that hypotheses involving linear dose responses with no threshold are untenable. The reader will have to determine if he or she can agree with the model Bond provides.

The next two papers deal with the use of somatic mutation assays at two different genetic loci. Albertini provides a concentrated review of the X-linked hprt gene, hypoxanthine-guanine phosphoribosyltransferase gene, which has been under investigation for over two decades. Jensen reviews the more recently developed autosomal glycophorin A (MN) blood group locus, which has greater relevance to the A-bomb survivor study.

The hprt gene mutation or variant is scored in T lymphocytes by two methods, a short-term assay that defines variant frequencies, untestable at the DNA level, and mutant frequencies determined by clonal assays. The system has been used to monitor humans for environmental exposures, smoking, chemicals, and radiation. Unfortunately, the persistence of mutants induced in adults is only three to four years at most because the mutant T lymphocytes in this case are not derived from the stem cell compartment but from the peripheral lymphoid tissue. Nevertheless, extensive molecular analysis has revealed considerable differences in the spectra of spontaneous and radiation-induced events. It is hoped that different mutagens may give different and recognizable spectra, and some data suggests such characterizations will be possible. Albertini suggests in a very intriguing discussion that hprt mutations of specific nonrandom DNA break sequences serve as surrogates for certain lymphoid malignancies because of identical DNA sequences.

These events are tied to illegitimate V(D)J recombinase activity responsible for the enormous immune system repertoire. These rearrangements at hprt are found at high frequencies in newborn and young children's blood, paralleling the childhood lymphoid cancer changes. He expects that there will soon be a myeloid hprt screening system available that will allow assessing mechanisms underlying adult hematopoietic malignancy.

Unlike the hprt system, which can be measured in any individual, the glycophorin system is applicable only to MN heterozygotes, namely about 50% of the population. Electronic screening systems can detect loss of either the M or N antigen expression in red blood cells which normally carry both and are fluorescently stained. Jensen describes the evolution of the assay system and its application to the A-bomb population, wherein a linear dose response was demonstrated in a subset of the population. One clear advantage then is that this red blood cell screening system has memory similar to the cytogenetic analysis; variants induced in the erythroid stem cell of the marrow are recovered in the mature RBC progeny. But such cells lack a nucleus, and the molecular nature of the event is not determinable in the recovered cell variant. Based on recent studies in Chernobyl, dose rate effects have been distinguished, i.e., much lower rates among chronically exposed versus acutely exposed individuals. It is noteworthy that at RERF the GPA assay has larger individual variation per unit dose than the cytogenetic assay, and given the dose-rate observation just cited, leads to the suggestion that the observed linear response in A-bomb survivors may be masking a linear-quadratic response.

In the final paper of this volume, Zimbrick discusses the future of the Radiation Effects Research Foundation (RERF). Taking up the issue of RERF's history, he summarizes the recent events that led the RERF visiting Japanese-American Scientific Council to recommend in June 1995 the establishment of a blue ribbon panel of experts from the international community to evaluate the scientific program. That committee met in February 1996 and issued its final report in June 1996 to both governments. Zimbrick also discusses the establishment of the two-year US Department of Energy contract with the National Academy of Sciences to continue its management role of RERF for the US side and remain a buffer against direct government intrusion.

With respect to future research programs, Zimbrick cites continued surveillance of the in utero exposed population regarding their potentially greater radiosensitivity than those exposed as adults, the need to assess more accurately the evidence for increased noncancer mortality risks, particularly through clinical studies of the Adult Health Study program, wherein confounding factors can be assessed. In addition to the first generation studies described in Neel's chapter, Zimbrick mentions the need to establish a well-designed clinical study on a subset of this group to determine the possible onset of late-acting detrimental mutations. He describes the enormous reservoir of stored biological samples at RERF and their importance to future studies using biomarkers to assess disease onset, mutational changes in tumor tissue, radiation repair capacity, dose-response relationships, and

other mechanistic issues. Finally he describes RERF's role in international collaboration with countries of the former Soviet Union that have also experienced radiation catastrophes.

In the five decades that have passed since the bombings of Hiroshima and Nagasaki, a number of scientific and economic developments have taken place. On a scientific note, there has been an explosive growth of new technologies related to the molecular characterization of disease and exposure. This has resulted in a virtual avalanche of knowledge on the role of exogenous and endogenous factors in the development of radiation-induced health effects. Not surprisingly, new findings from epidemiologic research cannot keep up with advancements being made in molecular-based work. On another note, economic and political changes have had significant impact on the approach of the Radiation Effects Research Foundation, the US Department of Energy, and the National Academy of Sciences toward the joint US-Japan program.

The present volume assesses the five decades of research on the survivors and their children and focuses on proceedings of the first Schull Symposium. Five broad areas of interest and activity were addressed: physics, cancer statistics and cancer epidemiology, genetics, molecular biology, and psychosocial effects and social responsibility. To reflect the focus of the symposium, the present volume is entitled Effects of Ionizing Radiation: Atomic Bomb Survivors and Their Children (1945–1995). Emphasis is placed on unifying the individual disciplines that serve to condition and shape the current outlook on human exposure to ionizing radiation. The current text devotes much more attention than earlier texts to the issues and problems faced by investigators in the early years of the Atomic Bomb Casualty Commission (ABCC), especially those relating to the National Academy activity abroad. Changes in international economic issues are also given broad coverage in the text.

The main themes of contemporary research are presented specifically in each of the chapters. The premise that exploration and analyses in radiation research are more meaningful when discussed in a historical context is reflected in the chapters as each author used a common thread to bridge existing gaps in knowledge. The successes and failures experienced by the contributors in their research over the past fifty years have had much bearing on reactions to challenges of the scientific and economic environments prevailing in the world today.

An international conference conceived and executed within twelve months owes a debt of gratitude to a number of individuals who made countless contributions to its success. The list is long but warrants inclusion.

Institutional sponsors include the Consulate General of Japan at Houston; the Atomic Energy Control Board of Canada; Offices of the President and Vice President of the University of Texas Health Science Center at Houston; the Executive Vice President for Administration and Finance of the University of Texas Health Science Center at Houston; the School of Public Health, Graduate School of Biomedical Sciences, and the Medical School of the University of Texas Health

Science Center at Houston; The Methodist Hospital; Baylor College of Medicine; the Department of Medicine at the University of Texas Branch at Galveston (Division of Oncology, General Medicine, and Office of Texas Department of Criminal Justice Affairs). The contribution from Dr. Joe Goldman and the International Center for Solutions of Environmental Problems provided not only office space and support but immeasurable moral support and guidance for the local organizing group of academic and community volunteers.

The academic committee consisted of Dr. Richard Wainerdi, Dr. Antonio Gotto, Dr. Jim Williamson, Dr. R. Palmer Beasley, Dr. Bill Butcher, and Dr. Don Powell.

Student scholarships for conference attendance were provided by Mr. and Mrs. Paul Bertin, Judith Booker, Dr. James Crow, Dr. Steve Daiger, Dr. Tommy Douglas, Tyrell Flawn, Dr. and Mrs. Arthur Garson, Dr. Lu-yu Hwang, Dr. David Hewett-Emmett, Mr. and Mrs. Joseph Meyer, Dr. Masotoshi Nei, David and Margaret Noble, Dr. Leif Peterson, Dr. Anthony Pisciotta, Melva Ramsay, Catherine Roberts, Mr. and Mrs. John Sellingsloh, Dr. Emoke Szathmary, Dr. K. Okamoto Dr. T. Aoyama.

Founding contributions to the conference were Ms. Sara Barton, Mr. Terry Bertin, Dr. Patricia Buffler, Dr. Ranajit Chakraborty, Dr. Darwin Labarthe, Dr. Don Powell, and Dr. Kim Dunn.

Community volunteers from the Learning Center for Sustainable Living and Foundation for Global Community were coordinated by Mrs. Catherine Roberts. The volunteers included Vickie Ratello, Carole Breckbill, Jim and Kate Conlan, Christine Economides, Don and Sharon Hill, Francis Jones, Mia and Victor Lamanuzzi, Eileen McGovern, Charles Orelup, Katherine Prelat, Barbara Stein, Bill and Majorie Tracy, and Yumi Yonehara. Students volunteers included Lisa Amelse, Mike Badzioch, Molly Bray, Elizabeth Bruckhoimer, Elena Capsuto, Dawn Chandler, Charles Earley, Karen Earley, Margaret French, Rob Harson, Julia Krushkal, Chun Hsin Lin, Grier Page, Brinda Raua, Wen Shui, Allison Stock, and Paul Wong.

The steering committee provided invaluable direction. It was composed of Dr. Phillip McCarthy, T.J. Dunlap, Armin Weinberg, Geraldine Gill, Paula Knudson, Besty Chadderdon, Carolyn Milton, and Melva Ramsay. Dr. Fred Tuthill provided legal counsel. Mr. Spencer Knapp, Amersham Corporation, is gratefully acknowledged for supporting the publication of this proceedings.

The planning group, without whose efforts this conference would not have occurred, consisted of Margaret Dybala, Deb Hall, Margaret Irwin, Catherine Roberts, Amelia Kurth, Terry Bertin, Sara Barton, and Dr. Kim Dunn.