Work on this paper was supported by NICHD Grant Nos. 5R01 HD-19915-02 and 5P30 HD-10003- 09. We are grateful to Robert W. Campbell, Robert A. Lewis, James R. Millar, William G. Rosenberg, and Lewis H. Siegelbaum for helpful comments on an earlier draft.

1. Steven Rosefielde, “The First ‘Great Leap Forward’ Reconsidered: Lessons of Solzhenitsyn'sGulag Archipelago,” Slavic Review, 39, no. 4 (December 1980): 559–87; “Reply,” ibid., pp. 612–15; “An Assessment of the Sources and Uses of Gulag Forced Labour 1929–56,” Soviet Studies, 23(January 1981): 51–87; “Reply to Nove,” Slavic Review, 41, no. 4 (Winter 1982): 766–71; “ExcessMortality in the Soviet Union: A Reconsideration of the Demographic Consequences of ForcedIndustrialization, 1929–1949,” Soviet Studies, 35 (July 1983): 385–409; “Excess CollectivizationDeaths, 1929–1933: New Demographic Evidence,” Slavic Review, 43, no. 1 (Spring 1984): 83–88.

2. R. W. Davies and S. G. Wheatcroft, “Steven Rosefielde's Kliukva,” Slavic Review, 39, no. 4(December 1980): 593–602; Stephen G. Wheatcroft, “On Assessing the Size of Forced ConcentrationLabour in the Soviet Union, 1929–56,” Soviet Studies, 33 (April 1981): 265–95; “Towards a ThoroughAnalysis of Soviet Forced Labour Statistics,” ibid., 35 (April 1983): 223–37; “A Note on StevenRosefielde's Calculations of Excess Mortality in the USSR, 1929–1949,” ibid., 36 (April 1984): 277–81.

3. Others who have commented on Rosefielde's interpretation of the role of forced labor inSoviet economic development include Igor Birman, “Limits of Economic Measurement,” Slavic Review, 39, no. 4 (October 1980): 603–607; Paul Gregory, “Valueless Goods and Social Bads in theMeasurement of Soviet Output Series, 1928–1932,” ibid., pp. 608–11; Holland Hunter, “The EconomicCosts of the GULag Archipelago,” ibid., pp. 599–92; Alec Nove, “Letter to the Editor, “Slavic Review, 40, no. 4 (Winter 1981): 691–93, and ibid., 41, no. 4 (Winter 1982): 771–73.

4. Conquest, Robert, The Great Terror (New York: Macmillan, 1968)Google Scholar; “Forced Labour Statistics:Some Comments,” Soviet Studies, 34 (July 1982): 434–39; Robert Conquest, Dana Dalrymple, JamesMace, and Novak, Michael, The Man-Made Famine in Ukraine (Washington, D.C.: American EnterpriseInstitute, 1984).Google Scholar

5. Mace, James E., Communism and the Dilemmas of National Liberation: National Communismin Soviet Ukraine, 1918–1933 (Cambridge, Mass.: Harvard University Press, 1983), pp. 280–96Google Scholar; and “Famine and Nationalism in Soviet Ukraine,” Problems of Communim, 33 (May-June 1984): 37–50.

6. Maksudov (pseud.), “Pertes subies par la population de l'URSS, 1918–1958,” Cahiers dumonde russe et sovietique, 18 (July-September 1977): 223–62; Dyadkin, Losif G., Unnatural Deathsin the USSR, 1928–1954 (New Brunswick, N.J.: Transaction Books, 1983).Google Scholar

7. Lorimer, Frank, The Population of the Soviet Union (Geneva: League of Nations, 1946).Google Scholar

8. Notestein, Frank et al., Future Population of Europe and the Soviet Union (Princeton, N.J.:Office of Population Research, “League of Nations Publications,” 1944).Google Scholar

9. Nicholas S. Timasheff, “The Postwar Population of the Soviet Union,” American Journal ofSociology, 54 (September 1948): 148–55.

10. Warren W Eason, Soviet Manpower: The Population and Labor Force of the U.S.S.R. (Ph.D.diss, Columbia University, 1959); “The Soviet Population Today,” Foreign Affairs, 37 (July 1959):598–606; “Labor Force,” in Abram Bergson and Kuznets, Simon, eds., Economic Trends in the SovietUnion (Cambridge, Mass.: Harvard University Press, 1963), pp. 38–93.Google Scholar

11. Conquest, “Forced Labour Statistics,” p. 438.

12. Lorimer, Population of the Soviet Union, p. 137.

13. Ibid., pp. 231–36.

14. Maksudov, “Geografiia goloda 1933 goda,” SSR: Vnutrennie protivorechiia, 1983, no. 7,pp. 5–17.

15. Rosefielde, “Excess Mortality,” p. 392.

16. Wheatcroft, “A Note on Steven Rosefielde's Calculations,” p. 279.

17. Coale, Ansley J., Anderson, Barbara A., and Harm, Erna, Human Fertility in Russia sincethe Nineteenth Century (Princeton, N.J.: Princeton University Press, 1979), pp. 16–17, 43, 199–206Google Scholar. A number of related social changes occurred in the 1930s, including rapid urbanization and separationof families, which could have affected the normal course of fertility.

18. Barbara A. Anderson and Brian D. Silver, “Estimating Census Undercount from School-Enrollment Data: An Application to the Soviet Censuses of 1959 and 1970,” Demography, 22 (May1985): 298–308.

19. Coale, Anderson, Harm, Human Fertility in Russia, p. 239.

20. Lorimer, The Population of the USSR, pp. 113–19.

21. Rosefielde, “Excess Collectivization Deaths,” p. 87.

22. Compare the original life tables calculated by S. A. Novosel'skii and V. V Paevskii, Smertnost'i prodolzhitel'nost’ zhizni naseleniia SSSR, 1926–1927. Tablitsy smertnosti (Moscow-Leningrad:Plankhozgiz, 1930) with the tables-reported in A. A. Vishnevskii and A. G. Volkov, eds., Vosproizvodstvonaseleniia SSSR (Moscow: Finansy i statistika, 1983), p. 298.

23. Ibid., p. 84.

24. Rosefielde, “Excess Mortality,” p. 390.

25. Ibid., p. 389.

26. Boris Urlanis, “Dinamika urovnia rozhdaemosti v SSSR za gody sovetskoi vlasti,” in A. G.Vishnevskii, ed., Brachnost', rozhdaemost', smertnost’ v Rossii i v SSSR: sbornik statei (Moscow:Statistika, 1977), pp. 11–12.

27. For example, Conquest, “Forced Labour Statistics “; Conquest et al., Man-Made Famine;and Mace, “Famine in Ukraine. “

28. Conquest, “Forced Labour Statistics,” p. 436; and Conquest et al., Man-Made Famine, p. 7.

29. For an evaluation of the quality and uses of Soviet census data, see Clem, Ralph S., ed.,Research Guide to the Russian and Soviet Censuses (Ithaca, N.Y.: Cornell University Press, forthcoming).CrossRefGoogle Scholar

30. Lorimer, Population of the Soviet Union, pp. 42 and 88. Lorimer notes a high ratio ofadolescent girls to boys and a deficit of young children in the 1926 census, and he speculates aboutthe reasons for these patterns. We think these patterns are consistent with those found in the 1959and 1970 Soviet censuses: undercounting of young children and of geographically mobile adolescents.

31. Anderson and Silver, “Estimating Census Undercount.” Underenumeration of young childrenin the 1959 census has also been noted by Godfrey Baldwin, Estimates and Projections of thePopulation of the U.S.S.R., By Age and Sex: 1950–2000, U.S. Department of Commerce, InternationalPopulation Reports, Series P–-91, no. 23 (Washington, D.C.: Government Printing Office,1973), p. 6.

32. Antonov-Ovseenko, Anton, Portret tirana (New York: Khronika, 1980), p. 211.Google Scholar

33. Using the 1959 census age-sex distribution, as Rosefielde has done, in an effort to validateAntonov-Ovseenko's 1937 population total is extremely problematic because it requires heroic assumptionsabout the course of fertility, mortality, and migration (annexation and displacement ofpopulations) during the period 1938–1958. This is why Rosefielde's attempt at such validation (see “Excess Mortality “) fails. For further discussion, see Wheatcroft, “A Note. “

34. The coverage of both the 1926 and 1939 censuses needs further study. The 1926 censusshould not be considered a completely accurate baseline. It suffered not only from the type ofmisenumeration by age noted in table 1, which is typically a symptom of other kinds of misenumeration,but probably also from underenumeration, especially of children and of mobile adolescentsand young adults. The cohorts who were adolescents and young adults in the 1926 census might havebeen counted more completely in 1939, so that there would be an apparent “excess” of people inthe age range 25 to 39 compared to what might be predicted on the basis of expected mortalitybetween 1926 and 1939. One should not be too hasty to conclude that any such excess is a sign ofpadding of the 1939 census count.

35. Rosefielde, “Excess Mortality. “

36. Rosefielde, “An Assessment.” When Eason calculated the difference between the numberof people working for pay in the 1939 census and the sum of such workers from economic establishmentand sector statistics, he found that the census exceeded the establishment and sector total by6.3 million. (See Eason, “Labor Force,” pp. 83–85.) Although Eason attributed this discrepancy tothe possible combination of a number of causes, including forced labor or other exclusions fromestablishment data, Rosefielde attributes not only the entire discrepancy of 6.3 million to forcedlabor but also the difference of 1.4 million between the 1939 census figure for the military andEason's judgment that the military in January 1939 could not have numbered more than 2 million.Wheatcroft has pointed out that more complete 1939 census data published along with the 1959census results eliminate the disparity that Eason had observed between the census count and theestablishment data for 1939. See “On Assessing the Size,” pp. 278–83.

37. USSR, Ts.S.U., Naselenie SSSR (chislennost', sostav i dvizhenie naseleniia) 1973: statisticheskiisbornik (Moscow: Statistika, 1975), p. 139. Life tables by sex for these years can be foundin Vishnevskii and Volkov, Vosproizvodstvo naseleniia SSSR, pp. 298–99.

38. For mortality schedules, we use Coale-Demeny East Model life tables (levels 9, ll^and13). We use model mortality schedules to obtain a range of mortality assumptions that spans thebest estimates of normal mortality schedules for the USSR during this period. Each level of themodel life tables provides a separate male and female mortality schedule, so that differential mortalityby sex is assumed.

39. Jean-Noel Biraben, “Naissances et répartition par age dans l'Empire russe et en Unionsoviétique,” Population, 31 (March-April 1976): 441–78. Biraben smoothed the age distribution toreduce the effects of age heaping. To help in making comparisons across censuses, Biraben alsoadjusted the 1926 Soviet population to correspond to the 1939 post-annexation borders. While maintainingthe proportions by single-year by sex estimated by Biraben, we scale down the total to matchthat reported in the 1926 census.

40. An age distribution for 1939 was published in the 1959 census (USSR, Ts.S.U., Itogi vsesoiuznoiperepisi naseleniia 1959 goda, SSSR: Svodnyi torn [Moscow: Gosstatizdat, 1962], p. 49).This distribution was for the post-annexation population, was not given separately by sex, and doesnot allow us to break the population into those aged 0–11 and 12+. However, Kolosova providessex ratios by age from the 1939 census (G. Kolosova, “Pol, vozrast’ i sostoianie v brake naseleniiaSSSR,” in G. M. Maksimov, ed., Vsesoiuznaiaperepis’ naseleniia 1970 goda: Sbornik statei [Moscow:Statistika, 1976], p. 169). And in 1956 the Central Statistical Board published an age distribution(not by sex) from the 1939 census for the pre-annexation population (see Vestnik statistiki, 1956,no. 6). This distribution provides a break between ages 11 and 12. Using these data in combinationwith Kolosova's data, it is possible to estimate the distribution of the pre-annexation 1939 populationby age and sex. Note that the Ts.S.U. apparently assumed that the age distribution of the annexedpopulation in 1939 was nearly identical to that for the Soviet population in pre-annexation boundaries. This can be inferred from the distribution of population by age in its reported pre- and postannexationfigures.

41. Vishnevskii and Volkov, Vosproizvodstvo naseleniia SSSR, p. 257.

42. USSR, Ts.S.U., Naselenie SSSR, p. 136.

43. This is true if one uses Rosefielde's estimates of the number of births in 1933 and 1934,which rely on his estimates of the population and Urlanis's estimates of the crude birth rates in thoseyears ( “Excess Mortality,” p. 388). Using our cohort-component method, deflating the estimatedbirth rates by 10 percent also decreases the number of projected births by over 1 million for thosetwo years.

44. Mace, Communism and the Dilemmas.

45. Mace, “Famine in Ukraine,” pp. 38–39; and Conquest et al., Man-Made Famine, p. 7.

46. Mace, “Famine in Ukraine,” p. 39.

47. Ibid., p. 38.

48. Ibid., p. 39.

49. Barbara A. Anderson and Brian D. Silver, “Estimating Russification of Ethnic IdentityAmong Non-Russians in the USSR,” Demography, 20 (November 1983): 461–90. On the assimilationof Ukrainians by Russians between 1926 and 1959, see Robert A. Lewis, Richard H. Rowland, andRalph S. Clem, Nationality and Population Change in Russia and the USSR: An Evaluation of CensusData, 1897–1970 (New York: Praeger, 1976), pp. 278–88 and 295–96.

50. Brian D. Silver, “The Ethnic and Language Dimensions in Russian and Soviet Censuses, “Brown University Population Studies and Training Center, Working Papers, no. WP-84–07 (August1984); forthcoming in Clem, Research Guide to the Russian and Soviet Censuses.

51. USSR, Ts.S.U., Slovari natsional'nostei i iazykov—dlia shifrovki otvetov na 7 i 8 voprosyperepisnykh listov (O natsional nosti, rodnom i drugom iazyke narodov SSSR) Vsesoiuznoi perepisinaseleniia 1970 g. (Moscow: Statistika, 1969). The analogous glossary for the 1959 census did notlist the Cossacks, either as a separate nationality or as a subgroup of another nationality. See USSR,Ts.S.U., Slovari natsional'nostei i iazykov—dlia shifrovki otvetov na 7 i 8 voprosy perepis'nogo lista(O natsional'nosti i rodnom iazyke) (Moscow: Gosstatizdat, 1959).

52. Wixman, Ronald, The Peoples of the USSR: An Ethnographic Handbook (Armonk, N.Y.:M. E. Sharpe, 1984), p. 208.CrossRefGoogle Scholar

53. S. I. Bruk and V I. Kozlov, “Etnograficheskaia nauka i perepis’ naseleniia 1970 goda, “Sovetskaia etnografiia, 1967, no. 5, p. 6.

54. Isupov, A. A., Natsional'nyi sostav naseleniia SSSR (Po itogam perepisi 1959 goda) (Moscow:Statistika, 1964), p. 12.Google Scholar

55. Lorimer, Population of the Soviet Union, pp. 137–40.

56. After this article went to press, in response to a challenge by Wheatcroft Mace defendedhis estimates of losses in the famine in the Ukraine (see “Ukrainian Famine,” Problems of Communism, 34 [March-April 1985]: 132–38). Mace concluded that an “irreducible minimum” of 5.5million Ukrainians died in the famine. But this estimate, like Mace's earlier ones, is of a populationdeficit, not excess deaths. The figure depends on assumptions about the rates of both mortality andfertility among Ukrainians for all of the nonfamine years between 1927 and 1938. Much of the “irreducible minimum” difference could result from an overestimate of the number of births or anunderestimate of the number of deaths in the nonfamine years.

57. Lorimer, Population of the Soviet Union, p. 240.

58. Rosefielde, “Excess Mortality,” p. 404.

59. Ibid.