Fig 1.
Annual strontium-90 deposition in the northern hemisphere from atmospheric nuclear testing.
Created by the author from UNSCEAR 2000 [2] Annex C, Table 7.
Fig 2.
Panel A: Trend in infant mortality rates in the U.S. The dashed line shows the predicted undisturbed trend. Panel B: Excess infant mortality rates, i.e. observed- minus predicted rates and the result of regression with Model (1) (bold grey line). The thin dotted lines show the three bell-shaped excess terms. The vertical lines indicate the time window used to determine the number of excess infant deaths.
Fig 3.
Deviations between observed and predicted infant mortality rates in units of standard deviations (standardized residuals) and range of ±2 standard deviations (broken lines).
Table 1.
Regression results for the United States with two competing trend models.
Table 2.
Improvement in fit to the U.S. data by stepwise model refinement.
Fig 4.
Panel A: Infant mortality rates in EU5 and predicted undisturbed trend (broken line). Panel B: Excess infant mortality rates, i.e. observed minus predicted rates and regression line.
Fig 5.
Deviations between observed and predicted infant mortality rates in units of standard deviations (standardized residuals) and range of ±2 standard deviations (broken lines).
Table 3.
Regression results for the pooled five European countries (EU5), 1950–2018.
Table 4.
Improvement in fit to the EU5 data by stepwise model refinement.
Table 5.
Regression results for the three central European countries, 1950–2018.
Table 6.
Regression results for the two southern European countries and EU5, 1950–2018.
Table 7.
Observed (O) and predicted (E) infant deaths in five European countries, 1950–2000.
Fig 6.
Panel A: Infant mortality rates in central- and southern Europe and predicted undisturbed trends (broken lines). Panel B: Excess infant mortality rates and regression lines.
Table 8.
Results of regressions with Model (1) of the data from central- and southern Europe.
Table 9.
Improvement in fit to the UK data by stepwise model refinement.
Fig 7.
Panel A: Excess infant mortality rates in the UK and regression line. The thin dotted lines show the five excess terms. Panel B: Standardized residuals. The broken vertical line indicates 1986, the year of the Chernobyl accident.
Fig 8.
Panel A: Infant mortality rates in EU5, 1931–2018. The interrupted line shows the unperturbed secular trend; the solid line shows the predicted trend adjusted for the effects of WW2. Panel B: Excess infant mortality rates and regression line.
Fig 9.
Observed minus fitted infant mortality rates in EU5 in units of standard deviations (standardized residuals) and the range of ±2 standard deviations (broken lines).
Table 10.
Observed and predicted infant deaths in 1950–2000 from regressions with Model (3).
Fig 10.
Panel A: Stillbirth rates in Germany, semilogarithmic plot, and reduced model (broken line). Panel B: Excess stillbirth rates and regression line.
Fig 11.
Standardized residuals from regression of stillbirth rates in Germany and range of ±2 standard deviations (broken lines).
Table 11.
Estimated timing (calendar year) of the two main infant mortality peaks.