Tall Tales: Anthropometric Measures of Well-Being in Imperial Russia and the Soviet Union, 1821-1960

К истории дискуссии.

На очевидные ошибки Б. Н. Миронова указывал еще в 1999 году Стивен Хок

Steven L. Hoch

TALL TALES: ANTHROPOMETRIC MEASURES OF WELL-BEING IN IMPERIAL RUSSIA AND THE SOVIET UNION, 1821-1960

Slavic Review 58, no. 1 (Spring 1999)

"From the soles of his feet to the roots of his hair, Paul Bunyan split the atmosphere exactly 12 feet 11 inches. His weight, he told me—and I don't doubt his word for a moment—was 888 pounds."¹ Together with precise measures of Paul Bunyan's hips, waist, shoulders, thighs, calves, and reach, subsequently detailed, we would, it seems, have a rather comprehensive anthropometric description of this legendary logger. But his companion, Babe the Blue Ox, "was seven axe handles wide between the eyes," quite a different anthropometric measure indeed. Moreover, "some persons give the measurement as forty-two axe handles and a plug of Star tobacco." And, we are told, "both figures are correct," all of which quickly brings us to the heart of the problem, species aside.²

Anthropometry, an important technique for the study of human aux-ology or growth, dates back to the eighteenth century, though it experienced long periods of scientific neglect. James M. Tanner revived this technique in the mid-1950s,³ and over a decade later, a small but prominent group of historians, including Emmanuel Le Roy Ladurie and Robert Fogel, began to make use of available data on the stature or height and, on occasion, the weight of military recruits from the nineteenth and twentieth centuries to assess changes in well-being. In Fogel's research, "mean height at specific ages is used as a measure of standard of living" and is used to "supplement other evidence such as indexes of real wages, estimates of per-capita income, and measures of food consumption."⁴Others have used secular changes in stature to estimate national income; to study changes in the standard of living, as conventionally understood, or in a biological standard of living; to predict mortality; and to determine aggregate well-being, though all are clearly quite distinct measures.⁵ It is

1. Earl Clifton Beck, "Paul's Dinner," Legends of Paul Bunyan, comp. and ed. Harold W. Felton (New York, 1947), 34.

2. Ivan Benson, "The Big Blue Ox," Paul Bunyan and His Men (Rutland, Vt., 1955), 31.

3. James M. Tanner, "A Concise History of Growth Studies from Buffon to Boas," in Frank Falkner and J. M. Tanner, eds., Human Growth, vol. 3, Neurobiology and Nutrition (New York, 1979), 515-93.

4. Robert William Fogel, "Physical Growth as a Measure of Economic Well-Being of Populations: The Eighteenth and Nineteenth Centuries," in Frank Falkner and J. M. Tanner, eds., Human Growth: A Comprehensive Treatise, 2d ed., vol. 3, Methodology, Ecological, Genetic, and Nutritional Effects on Growth (New York, 1986), 263. See also John Komlos, Nutrition and Economic Development in the Eighteenth- Century Habsburg Monarchy: An Anthropometric History (Princeton, 1989), 23-54.

5. H.J. Brinkman, J. W. Drukker, and B. Slot, "Height and Income: A New Method for the Estimation of Historical National Income Series," Explorations in Economic History 25 (July 1988): 227-64; John Komlos, ed., The Biological Standard of Living on Three Continents:

Slavic Review 58, no. 1 (Spring 1999)

this same method and its underlying premises that are taken up in the research here.

Boris Mironov and Stephen G. Wheatcroft both relate observed changes in height, and in the latter case sometimes weight, to shifts in biological status, clearly a different measure of well-being than standard of living. Both recognize that a complex of factors affect biological status, although these considerations are not always consistently applied in Miro-nov's analysis. They adhere, in principle, to the well-established medical view that "the most important exogenous factors [non-genetic] influencing growth are nutrition and disease."⁶ As Phyllis Eveleth and James Tanner write: "There are many environmental factors which affect growth, and which combine in various proportions to constitute the environment of poverty; but in the final analysis most of them hinge upon the level of nutrition in conjunction with the prevalence of childhood infection."⁷

The use of anthropometric techniques to analyze time series shifts in the biological status of a population presents a number of problems. The issues relevant here include the validity of the actual measures used (do they measure what they say they measure), the accuracy of the recorded data, the periodization of the observed changes in the measures, and the causes of the observed changes, each of which will be discussed in turn. The importance of this discussion is to determine whether the application of anthropometric methods to available Russian and Soviet data can sustain the conclusions reached by Mironov and Wheatcroft.

Mironov relies solely on data on terminal height or stature (which is generally attained between the ages of 20-25) as an indicator of biological status, contending that "even the most insignificant increase or decrease in height can indicate a change in the biological status of a population." In contrast, Wheatcroft supplements his information on terminal stature whenever possible with available information on mortality, weight, weight-for-height or body mass indexes, height-for-age, and birth weights, reflecting many of the concerns medical professionals have regarding anthropometric measures, concerns that have more recently been directed to historians as well. Wheatcroft's approach is certainly more helpful because of the well-recognized problem of using a single variable, especially one that is widely regarded as imprecise, as we shall see.

Further Explorations in Anthropometric History (Boulder, Colo., 1995); Jane Humphries, "Short Stature among Coal-Mining Children: A Comment," Economic History Review 50, no. 3 (1997): 531-37; Roderick Floud, Kenneth W. Wachter, and Annabel Gregory, Height, Health, and History: Nutritional Status in the United Kingdom, 1750-1980 (Chicago, 1994). See also Robert W. Fogel, "Nutrition and the Decline of Mortality since 1700: Some Preliminary Findings," in Stanley L. Engerman and Robert E. Gallman, eds., Long Term Factors in American Growth (Chicago, 1986), 439-555.

6. C. Van Wieringen, "Secular Growth Changes," in Falkner and Tanner, eds., Human Growth: A Comprehensive Treatise, 3:322; and James R. Hummert and Alan H. Goodman, "An Assessment of the Effects of Childhood Illness on Growth in Height and Weight," Growth 50 (1986): 371.

7. Phyllis B. Eveleth and James M. Tanner, Worldwide Variation in Human Growth, 2d ed. (Cambridge, Eng., 1990), 191.

Early practitioners of anthropometric history focused on terminal stature or height, as it was often the only anthropometric information available. In most of Europe, males eligible for military service were measured for height at ages close to their expected terminal height, as was the case in imperial Russia and the Soviet Union. But the current state of anthropometric research will no longer support this approach. A 1991 study on the Anthropometric Assessment of Nutritional Status noted: "The most common anthropometric measures are weight and changes in weight, stature, segmental lengths, fat folds, and various body circumferences and diameters. . . . Their utility is that they provide rough estimates of body composition or changes in body composition if they are taken serially. They have the advantage of being inexpensive, easy to carry out, and reproducible with a modicum of training. Their main disadvantage is their lack of precision."⁸ Because of this high level of uncertainty, current anthropometric standards for the evaluation of growth require twenty-five different measures including stature; weight; body mass index; sitting height; bio-chanteric breadth; elbow breadth; mid-upper arm circumference, area, and muscle area; mid-arm fat; triceps and subscapular skinfold thickness; and percent fat weight.⁹

Following earlier lines of research, Wheatcroft states: "Data on terminal height provide a history of 'net nutrition' up to the age at which growth ceases (normally between 20 and 24). Weight is an indicator of more recent net nutritional status."¹⁰ Mironov holds that "an expert majority of anthropometricians at the present time endorse the paradigm that net consumption exerts a decisive influence on body height, and in this analysis, I rely on that paradigm." But it is unclear how useful or sensitive final stature alone is as a measurement of well-being, however defined. "The major increase in the prevalence of stunting . . . occurs in the first two years of life. After that period the curve flattens." Height-forage, weight-for-age, and weight-for-height as anthropometric measures are perhaps better applied to children only through the age of 7 "because from then on genetic factors start playing a greater role in the determination of a child's height and weight."¹¹ To quote Eveleth and Tanner again: "The average values of children's heights and weights reflect accurately the state of a nation's public health and the average nutritional status of the population of its citizens, when appropriate allowance is made for differences, if any, in genetic potential. This is especially so in developing or disintegrating countries."¹² For these same reasons, comparisons

8. Johanno T. Dwyer, "Concept of Nutritional Status and Its Measurement," in John H. Himes, ed., Anthropometric Assessment ofNutritional Status (New York, 1991), 16 (emphasis added).

9. A. Roberto Frisancho, Anthropometric Standards for the Assessment of Growth and Nutritional Status (Ann Arbor, 1990).

10. See also Komlos, Nutrition and Economic Development, 27.

11. Kees Kostermans, Assessing the Quality of Anthropometric Data (Washington, D.C., 1994), 5-6; WHO Working Group, "Use and Interpretation of Anthropometric Indicators of Nutritional Status," Bulletin of the World Health Organization 64, no. 6 (1986): 931.

12. Eveleth and Tanner, Worldwide Variation, 1 (emphasis added).

with populations of differing genetic background are impossible, though Mironov attempts this. Even current research admits the difficulty of establishing nonpopulation-specific anthropometric reference standards that can permit such comparisons.¹³ Equally, the methods employed require the assumption that the genetic composition of the population measured across the period under observation remain relatively stable, proportionally speaking. This is true for only some of the populations or subgroups analyzed by Wheatcroft and Mironov.

Although data concerning final stature are unfortunately often the only data available to historians, the dependability of this measure as an indicator of well-being is low.¹⁴ Given the problems of relying exclusively on stature or height, the World Health Organization noted as early as 1986: "It is generally desirable to use more than one method of presenting and analyzing anthropometric data."¹⁵ Heeding this caution, John Riley, using data on height and weight to compute a body mass index for adult males in Britain in the 1860s rather than relying on height alone, demonstrated, as suggested by the World Health Organization, that "weight matters. The addition of information about weight, which allows calculation of body mass, calls into question the use of height as a gauge of quality of life, nutritional status, and mortality risk."¹⁶ Similarly, recent research on men of Amherst College between 1834-1949 found that "no relationship between height and mortality could be detected," while at the same time a "bulk-mortality relationship received strong confirmation."¹⁷ Examination of the height and health of the army of the East India Company has led to the conclusion that "it may be time to tone down some of the overly hopeful claims made for the application of anthropometric data to economic history made in the earlier stages of the literature."¹⁸ In all, it would seem that height without weight cannot bear the burden of anthropometric analysis.

Another factor confounding the use of anthropometric measures, and especially of height as a measure of well-being, is the autonomous role of infectious disease. During the period under consideration, imperial Russia and the Soviet Union experienced major changes in the infectious disease environment. Smallpox was brought under control in many parts of the empire by the end of the nineteenth century and eradicated, in all likelihood, just prior to World War II. Cholera first became epidemic in the 1830s; poliomyelitis in the 1950s. Tuberculosis became endemic to-

13. Frisancho, Anthropometric Standards, 3.

14. Kostermans, Assessing the Quality, 3.

15. WHO Working Group, "Use and Interpretation," 938.

16. John C. Riley, "Height, Nutrition, and Mortality Risk Reconsidered," Journal of Interdisciplinary History 24, no. 3 (Winter 1994): 492.

17. John E. Murray, "Standards of the Present for People of the Past: Height, Weight, and Mortality among Men of Amherst College, 1834-1949, "Journal of Economic History 57, no. 3 (September 1997): 585.

18. Joel Mokyr and Cormac O'Grada, "Height and Health in the United Kingdom 1815-1860: Evidence from the East India Company Army," Explorations in Economic History 33 (April 1996): 163.

ward the end of the last century. Measles and whooping cough may well have become more prevalent, and scarlet fever less virulent.¹⁹ In addition, a shift in breast-feeding practices and the increased availability of unadulterated milk supplies may well have reduced the biggest killer of all, infant and childhood diarrheal diseases, often implicated in stunting. All these diseases affect quite different age groups and have differential consequences for human growth.²⁰ Although some of these changes in the disease environment can be attributed to direct medical or public health interventions, others are due to random variation in the nature of disease or shifts in the seasonal demand for women's labor, all of which complicate any analysis of changes in well-being.

Finally, even to the degree that anthropometric measures are correlated to the net nutritional status or well-being of a population, it needs to be noted that an entitlement to food, especially on the part of the young, reflects social values, and not simple availability.²¹ This is especially the case in patriarchal societies in which wealth and status are stratified by generation. Changes in entitlement practices resulting from fundamental shifts in attitudes toward children can be a significant factor affecting nutritional status and well-being, quite apart from net gain or loss in availability.

The raw data used by Mironov and Wheatcroft present formidable problems for analysis, especially given that the 100-year secular increase in stature under discussion is less than 10 cm, or just under 4 inches (see Wheatcroft, figure 3). First, it is important to understand that Mironov, exclusively, and Wheatcroft, frequently, use summary data, grouped variously in anywhere from five to seventeen height intervals. Such an inconsistent grouping of the raw data across time can seriously distort serial estimates of mean stature. Second, they are often using truncated data, that

19. Kenneth F. Kiple, ed., The Cambridge World History of Human Disease (New York, 1993), 1063; Joseph H. Bates and William W. Stead, "Tuberculosis as a Global Epidemic," Medical Clinics of North America 77, no. 6 (November 1993): 1208-10; Steven L. Hoch, "Famine, Disease, and Mortality Patterns in the Parish of Borshevka, Russia, 1830-1912," Population Studies 52, no. 3 (October 1998); K. G. Vasil'ev and A. E. Segal, Istoriia epidemii v Rossii (Materialy i ocherki) (Moscow, 1960), 214-353; O. V. Baroian, Results of Half a Century of Combating Infections in the USSR and Certain Urgent Problems of Modern Epidemiology (Moscow, 1968); E. I. Chazov, ed., 70 let sovetskogo zdravookhraneniia, 1917-1987 (Moscow, 1987).

20. Hans-Joachim Voth and Timothy Leunig, "Did Smallpox Reduce Height? Stature and Standard of Living in London, 1770-1873," Economic History Review 49 (1996): 541-60; Ida Johnston et al., "Reading Attainment and Physical Development after Whooping Cough,'¹'Journal ojEpidemiology and Community Health 39 (1985): 314-19; R. C. Hauspie and H. Pagezy, "Longitudinal Study of Growth of African Babies: An Analysis of Seasonal Variations in the Average Growth Rate and the Effects of Infectious Diseases on Individual and Average Growth Patterns," Acta Paediatrica Scandinavia, Supplement 350 (1989): 37-43; A. Prader, J. M. Tanner, and G. A. Von Harnack, "Catch-up Growth Following Illness or Starvation," Journal of Pediatrics 62 (1963): 646-69; and Hummert and Goodman, "An Assessment," 371-77.

21. Richard H. Steckel, "Stature and Standard of Living," Journal ojEconomic Literature 33 (December 1995): 1904.

is, only for recruits who met the minimal height requirements, a level that changed over time. Third, problems such as age-heaping, frequently encountered in historical demography, and difficulties in obtaining accurate height measurements, cited in many anthropometric studies, could skew their height-for-age results. Fourth, no serial data appear to be available on changes in the proportion of men who fell below the minimum height requirement. This will distort any calculation of mean height in a population. Fifth, it is unclear that recruits, especially prior to 1874, are representative of the population as a whole. Landlords and peasant communes with good reason often selected recruits from the poorest households.²²Finally, in much of the nineteenth century, annual recruitment levies were regional, and Mironov admits to regional differences in stature.

Though both authors admit to some of these problems, they do not appear able to correct them. They express serious reservations about the data, but proceed anyway. Neither Wheatcroft or Mironov appear to have undertaken any internal tests on the data to see if they conform to known distributions of stature in Human populations. It is widely recognized that height reflects a Gaussian distribution, and it would have been useful to know if the summary data employed conform to this anticipated distribution.²³ This would have permitted Mironov to test for accuracy over time and to correct for shifts in the minimal height requirements for recruitment. Finally, no evidence is presented in Wheatcroft's essay that will allow me to agree with his claim that "Soviet statistical data are generally far more reliable than are normally presumed." Wheatcroft candidly admits to problems of "censorship," "purposefully misleading data," "demonstrably false" claims, and "selectively reproduced figures." But no analysis of the data on stature, either statistical or substantive, is presented to demonstrate its reliability and justify its use. At present, I remain unconvinced.

For all these reasons, it is not surprising that Wheatcroft and Mironov come up with contradictory results for the years in which their research overlaps. Wheatcroft identifies "two fairly steady periods in which there was a distinct upward trend: 1) from the 1850s birth cohort to the late 1880s, and 2) from the mid-1930s birth cohort to the late 1950s." In contrast, Mironov holds, "for 1821 to 1910, we can discern three periods: in the first period, 1821-1850, or roughly thirty years, physical stature increased; in the second, the fifteen years between 1851 and 1865, stature decreased; and in the third, between 1866 and 1910, or forty-five years, stature again increased." Wheatcroft also identifies two major periods of decline "1) for birth cohorts 1899-1909, and 2) for birth cohorts 1923-1934, and possibly one minor decline for birth cohorts 1913-1918." In other words, between 1851 and 1865, Mironov has stature decreasing and Wheatcroft has it increasing; Wheatcroft sees this increase continuing to

22. Steven L. Hoch, Serfdom and Social Control in Russia: Petrovskoe, a Village in Tambov (Chicago, 1986), 155; Rodney D. Bohac, "The Mir and the Military Draft," Slavic Reviexu 47, no. 4 (Winter 1988): 652-66.

23. Eveleth and Tanner, Worldwide Variation, 5; Frisancho, Anthropometric Standards, 31.

the late 1880s, followed by a major decline in the first decade of the twentieth century. Mironov sees an increase from 1866-1910.

These discrepancies highlight the problems of trying to periodize observed changes in final stature with shifts in either biological well-being or standard of living. In organizing height data of 20-year-old recruits by birth cohort, Mironov is of the view that "the difference in the height of recruits between 1877 and 1876—that is, 1 mm—can essentially be explained by that one year, 1856." Consequently, Mironov contends that "the annual change in body height permits one to determine when the tendency toward increased height is replaced by the tendency toward decreased height, and correspondingly, when the tendency toward improved biological status is replaced by a tendency toward declining biological status, and vice versa." This is a serious statistical flaw that undermines the entire discussion that follows. In essence, Mironov uses cohort analysis to draw conclusions that can only be drawn from period analysis. Cohort (or longitudinal) analysis traces the fate of a specific group of individuals from the time the cohort is exposed to the risk being studied, in this instance the risk of being short in stature which begins at age 0 and lasts until age 20-25. In contrast, period (or cross-sectional) analysis involves "a quantitative description and analysis of the data for many cohorts observed during a specified time interval, such as a year, with respect to some variable."²⁴

In other words, the final stature of a birth cohort can be affected by factors across a considerable period of time and, thus, observed changes in height cannot readily be periodized or fixed in time. First, Tanner has noted that a "contributory cause of the shortness [persisting into adult life] might have been the undernutrition of mothers much earlier, when they themselves were children, causing stunting of size and in consequence the production of small babies."²⁵ Thus, we are presented with the possibility that a cohort of short mothers born in 1820, giving birth to children largely between 1840-1860, had male sons who, when measured between 1860-1880, were found to be of short stature.

Second, the impact of childhood infectious diseases on final stature is quite age-specific. In Russia, diarrheal diseases predominated in the very early months of life, as early weaning was common. Measles, smallpox, and whooping cough occurred slightly later. Diphtheria and scarlet fever primarily affected children aged 3-4. The age-specific nature of various infectious disease insults results in quite different impacts on successive birth cohorts. Thus, Mironov is incorrect in his assumption that the cause of stature differences in successive birth cohorts is limited to events occurring in the first year of life in the older cohort.

Third, and of greater import, as Wheatcroft recognizes, periodization is greatly confounded by the quite varied nature of catch-up growth. Some birth cohorts stunted at a younger age may well have had both the

24. Henry S. Shryock and Jacob S. Siegel, The Methods and Materials of Demography (NewYork, 1976), 550-51.

25. Tanner, "A Concise History of Growth Studies," 3:546.

time and the food energy to recover. Other birth cohorts may simply have lacked the time; others may have lacked sufficient nutrition to recover all of the growth lost. At any age, "it is difficult to determine if complete catch-up has occurred."²⁶ Moreover, there have been instances which "demonstrate that the fluctuations in the share of undersized conscripts generally coincided with the food situation in the years immediately preceding the examination."²⁷ Finally, tuberculosis, which first entered the disease environment during this period, inhibits catch-up growth far more than most other diseases, and thus would be a new random factor affecting final stature.²⁸

These problems make period analysis of all but very long-term secular changes impossible and call into question Mironov's detailed but unsup-portable periodization of shifts in biological status and all the conclusions that follow. For any one birth cohort, a complex of factors across a sixty-year period needs to be taken into consideration. Wheatcroft's basic argument does not rest on a detailed periodization of changes in final stature, though at times he does imply one. But, in fact, the factors influencing the final stature of a birth cohort are hard to determine and impossible to date with any degree of accuracy. Consequently, shifts in biological well-being cannot be determined with the chronological precision Wheatcroft and especially Mironov suggest. As Joel Mokyr and Cormac O'Grada have cautioned: "Time-series inferences from height data therefore remain, in our view, hazardous."²⁹

In reference to observed changes in stature, Mironov states: "The task of the historian is to explain which components of this [biological] status have produced these changes; these components would include the quality of nutrition, intensity of work or training, illness, environment, massive and prolonged stress, and physical overexertion." In contrast, Wheatcroft is much more cautious, generally eschewing any discussion of causality. He is content to observe the improved stature in the population in the long run, which he holds, serves as a counterweight to well-known periods of demographic and economic crisis and biological stress. Again, Wheatcroft's approach is more prudent.

To Massimo Livi-Bacci: "Variations in height reflect not only variations in nutrition but also differences in the frequency of contagious diseases, independently of prevailing nutritional habits. The state of the art does not yet enable us to unravel the causal tangle of food-hygiene-infection-height, even in current studies."³⁰ Similarly, attempts to weigh "the rela-

26. Jere D. Haas and Jean-Pierre Habicht, "Growth and Growth Charts in the Assessment of Pre-School Nutritional Status," in C. A. Harrison and J. C. Waterlow, eds., Diet and Disease in Traditional and Developing Societies (Cambridge, Eng., 1990), 175.

27. C. A. Mandemakers and J. L. Van Zanden, "The Height of Conscripts and National Income: Apparent Relations and Misconceptions," Explorations in Economic History 30 (January 1993): 83.

28. Haas and Habicht, "Growth and Growth Charts," 161.

29. Mokyr and O'Grada, "Height and Health," 163.

30. Massimo Livi-Bacci, Population and Nutrition: An Essay on European Demographic History, trans. Tania Croft-Murray, with the assistance of Carl Ipsen (Cambridge, Eng., 1990), 108.

tive individual importance of such variables on growth is further complicated by the complexity of their synergistic relationship. . . . Dietary deficits and disease cannot be completely disassociated when evaluating their effects on growth and development. This is undoubtedly true in underdeveloped countries in which diseases are potentially life threatening, generally show a profoundly negative effect on growth, and are frequently found in consort with various forms of malnutrition."³¹ In the end, Mironov discovers, as have many others, that shifts in stature do not correlate well with other indicators of well-being. Mironov blames the other variables. Given all the problems of anthropometric history, this does not appear to be justified.

Wheatcroft's research into stature highlights the fact that short-term crises do not preclude important secular improvements in well-being. In fact, recovery is often swift. This accords with the findings of Susan Cott Watkins and Jane Menken and of Ajit Das Gupta, as well as with my own findings.³² His suggestion that this secular gain in well-being occurred without an improvement in food availability implies that future lines of research should focus more intently on public health policies and campaigns as well as changes in the infectious disease environment. Equally significant, his description of a long-term increase in height provides additional support for established views. In a well-known collection of essays published in 1960 by some of America's best cold warriors, a long-term increase in life expectancy was recognized as the result of "in good part a genuine improvement in conditions affecting mortality, as may be seen in data on life expectancy at birth: 32 years in 1896-97, 44 years in 1926-27, 48 years in 1938-1940 and 67 years in 1955-1956," much attributable to a significant decline in infant mortality, as is to be expected.³³ This view has not changed in the last forty years. A 1999 edition of a textbook on Russian and Soviet history, in discussing the period up to 1960, states "the USSR was in the forefront of improving public health, raising life expectancy, and reducing infant mortality."³⁴ To quote Wheatcroft, "there were remarkable changes and improvements in welfare, mortality, and height."

On this basis, Wheatcroft urges us to reassess "what the Soviet experience meant to the mass of the Soviet population." He concludes, "it 'probably' was the best time for most of its citizens." Rebuking Richard Pipes, Zbigniew Brzezinski, and Robert Conquest, Wheatcroft prefers to stand in the camp of our legendary logger. "Paul Bunyan was the one historian of the useful and the beautiful; other writers of history tell only of terrible

31. Hummert and Goodman, "An Assessment," 371.

32. Susan Cott Watkins and Jane Menken, "Famines in Historical Perspectives," Population and Development Review 11, no. 4 (1985): 654; Ajit Das Gupta, "Catastrophes and Their Population Symptoms," in Hubert Charbonneau and Andre LaRose, eds., The Great Mortalities: Methodological Studies of Demographic Crises in the Past (Liege, 1980), 65; Hoch, "Famine, Disease, and Mortality Patterns," 363.

33. Warren W. Eason, "Population Changes," in Cyril E. Black, ed., The Transformation of Russian Society: Aspects of Social Change since 1861 (Cambridge, Mass., 1960), 79-80.

34. David MacKenzie and Michael W. Curran, A History of Russia, the Soviet Union and Beyond, 5th ed. (New York, 1999), 573.

and dramatic events."³⁵ To Wheatcroft, as to many before him, the Soviet experience is to be subjected to cost-benefit analysis. We need "to gain a more balanced view" as we debate "the achievements of Soviet communism, and how these should be weighed against the costs." In particular, he argues that short-term crises, however emotionally or morally charged, must be viewed in the perspective of significant secular improvement.

Social scientists and historians should indeed pay attention to what the Soviet regime accomplished. Markers of ordinary people's lives are useful and help illuminate the views and opinions of those who lived through this period. These achievements help us understand the social bases of support for the regime. But good and evil, certainly familiar terrain for the historian, cannot always be the subject of cost-benefit analysis. Even those regimes most would recognize as evil had social welfare goals for society, and clearly Soviet ideology demanded that. For the vast majority of Germans living between 1914 and 1950, the early years of Nazi rule were the best of times. But none of the accomplishments of the Nazi regime will permit me to view it in a more beneficial light.

Although it might be possible to use stature as an indicator of economic or biological well-being, we need to be cautious of rhetorical slips suggesting that we can measure intellectual and ethical health similarly. They cannot be placed in the same ledger book. The new information provided by Wheatcroft leads him to make moral claims. But improvements in stature or life expectancy must not be allowed to place the Soviet experience "into some form of perspective," rebalancing the moral ledger. Pipes, Brzezinski, and Conquest are not simply discussing mortality data as a measure of well-being, for which other indicators indeed might be more useful, they are describing state policies of extermination, and we must not confuse these two distinct aspects of life. It may be ironic that at a time of great human brutality there was a significant gain in economic or biological well-being. But such improvements cannot serve as a counterweight to the repressive policies of the Soviet state. Rather, we need to be reminded of Charles Dickens: "It was the best of times, it was the worst of times," an age of wisdom, foolishness; belief, incredulity; Light, Darkness; hope, despair; everything before us, nothing before us. All these were surely part of the Soviet experience.

35. James Stevens, "The Winter of the Blue Snow," in Paul Bunyan (New York, 1925), 11.