Contents - Previous - Next

Birth weights and stillbirths in historical perspective

Sources of evidence
Concluding remarks

RH Steckel

Correspondence: Dr Richard Steckel

Departments of Economics & Anthropology, Ohio State University, Columbus, OH 43210, USA

Data on birth weights and the percentage of stillborn babies during the late nineteenth and early twentieth century, taken from hospitals records in Edinburgh, Vienna, Dublin, Boston, Montreal and Philadelphia reflect the social class and living standard of their patients as well as socio-economic developments in these cites. To the extent that these hospital records also contained information on the mother's age, parity, occupation, ethnic and religious background, marital status and gender of the child, some inferences can be made on influences on intrauterine growth and birth outcome.


This workshop focuses on the consequences of intrauterine growth retardation (IUGR). The historical record reveals great variation in the prevalence and severity of IUGR as judged by the large differences in the birth weights of comparable populations over time. The adverse factors responsible are to be found in many different combinations in the immense variety of social and economic conditions under which people have lived in earlier times. The past can thus be a valuable laboratory for understanding the effects of these factors on growth and development. The limitations of historical data must also be recognized. They often do not meet modern criteria for their collection and the needed outcome measures may not have been recorded. Nevertheless, when properly qualified and interpreted, the historical record can shed light on antecedents and possible mechanisms of IUGR.

Sources of evidence

Although the accurate weighing of infants at birth can be traced to the mid-eighteenth century, historical evidence is not available in quantity until the second quarter of the nineteenth century when various European maternity hospitals routinely weighed newborns (Tanner, 1981; Ward, 1993). Several North American hospitals followed near the middle of the century. The early investigators were inspired by diverse motives, including intellectual curiosity, a desire to understand human biology, and knowledge of the relationship of size to survival of mother and child. During the second half of the century, some investigators regarded birth weight as a useful measure of fetal development, but debate and confusion over this concept persisted well into the twentieth century. By the end of the nineteenth century, however, weight and length at birth had become tools of pediatric diagnosis that were being used to deepen understanding of causes underlying variation in newborn size.

Among historians, W. Peter Ward (1993) has made the most substantial investigation of nineteenth and early twentieth century birth weights. He examined the weights of newborns at Edinburgh, Vienna, Dublin, Boston, and Montreal. In addition, Goldin and Margo (1989) have studied the hospital records for Philadelphia. These cities are examples of the many urban centers that established maternity hospitals by the late nineteenth century. The hospitals were usually supported by private philanthropy, and often by religious organizations or by the government. In addition to their charitable objectives, the hospitals facilitated the training of students and midwives and were useful to physicians in the study of reproduction, childbirth, and infant care.

Although these cities were centers of economic growth and change during the nineteenth century, they were also places of extremes in living standards that were related to social class and business cycles. The well-off typically gave birth in the home or in private maternity boarding houses, whereas the poor, the unfortunate, and the unmarried found a haven in the lying-in hospitals. With gains in knowledge and improvements in medical practices, the stigma of childbirth in maternity hospitals eventually declined during the early twentieth century, and these institutions began to draw patients from a wider spectrum of society. By the 1930s, a substantial share (perhaps a majority in some locations) of births took place in hospitals in many industrial countries.


Figures 1a through 5c present information on birth weights in Edinburgh, Vienna, Dublin, Boston, and Montreal during the nineteenth and early twentieth centuries. All data were collected and discussed by Ward (1993). The results are arranged in groups of three figures for each city. The first figure in each set presents average weight by year of birth for live-born singletons weighing 1500 grams or more. The second figure shows the per cent of live births of low birth weight (under 2500 grams), and the third figure gives the per cent stillborn.

The evidence for Edinburgh comes from a sample of 8,891 live births at the Royal Maternity Hospital, which was founded in 1844. A large majority of its maternity patients in this period were unmarried women involved in domestic service or unskilled occupations. Figure la shows considerable fluctuation in mean birth weights around an overall average of roughly 3,300 grams. Interestingly, average weights decreased approximately 400 grams in the last quarter of the nineteenth century and recovered somewhat in the early twentieth century. The deteriorating outcomes of pregnancy were mirrored in the time patterns in percentages of low birth weight and stillbirths. Remarkably, stillborns exceed 15 per cent of pregnancy outcomes in the early twentieth century.

The Vienna evidence on 10,111 live births was taken from the records of the Allgemeines Krankenhaus, where most patients were unmarried Roman Catholics. Figure 2a shows that weights in most time periods fluctuated around a mean of slightly more than 3,100 grams. There is a hint of deteriorating conditions prior to 1920 in the birth weights, which is more recognizable in percentages of low birth weight and stillbirths. The per cent of low birth weight increased from around 10 in the 1860s to more than 15 around World War I, while the per cent stillborn roughly doubled from 5 to 10.

The Rotunda, one of the oldest lying-in hospitals in Europe, is the source of data for Dublin. The patients in this sample were unusual for maternity hospitals in that only 5 per cent were single. As a result, more of the women had better socio-economic circumstances and they delivered higher order births, which tend to be heavier. Figure 3a shows a modest upward trend in weights around a mean of approximately 3,300 grams. The percentage of low birth weight declined in the last quarter of the nineteenth century from approximately 10 per cent to about 7 per cent in the early twentieth century. The incidence of stillborns fluctuated between 5 and 10 per cent but showed no obvious long-term trend. The Boston data were taken from the New England Hospital for Women and Children, which was located in an upwardly-mobile working class district, and the inpatient and outpatient services of the Boston Lying-in Hospital, located in an area that was an immigrant slum by the 1880s. Figure 4a makes clear the birth weight advantage of women served by the New England Hospital, where the overall mean was 3,480 grams. The relatively high overall mean was reflected in the incidence of low birth weight, which fluctuated in the area of 5 to 6 per cent, and in an incidence of stillbirths of approximately 2 to 5 per cent. Blacks, virtually all of whom were served by the inpatient service of the Lying-in Hospital, had birth weights that were about 200 grams below those of inpatient whites and 350 grams below those of outpatient whites and of New England Hospital patients.

Figure 1a. Edinburgh, 1847-1920: Mean Birth Weight

Figure 2a. Vienna, 1865-1930: Mean Birth Weight

Figure 1b. Edinburgh, 1847-1920: Percentage Low Birth Weight (3-year running average)

Figure 2b. Vienna, 1865-1930: Percentage Low Birth Weight (3-year running average)

Figure 1c. Edinburgh, 1847-1920: Percentage Stillborn (3-year running average)

Figure 2c. Vienna, 1865-1930: Percentage Stillborn (3-year running average)

The University Lying-in Hospital, founded in 1843, is the source of data for Montreal. This institution served predominately the English-speaking population. Overall, about 60 per cent of its patients were unmarried, but the share increased from about 50 per cent to 75 per cent over the time period observed. Many mothers in the early period were Irish immigrants but their numbers declined steadily such that the Canadian-born predominated after the 1860s. Figure 5a shows that birth weights near the middle of the century averaged about 3,500 grams, but then plummeted after 1870 to less than 3,100 grams at the turn of the century. The percentage of low birth weight was remarkably small (under 5 per cent) near the middle of the century and then rose to 10 per cent by 1900. The per cent stillborn rose from about 5 per cent in 1870 to about 12 per cent in 1900.

Figure 3a. Dublin, 1869-1930: Mean Birth Weight

Figure 4a. Boston, 1872-1900: Mean Birth Weight

Figure 3b. Dublin, 1869-1930: Percentage Low Birth Weight (3-year running average)

Figure 4b. Boston, 1872-1900: Percentage Low Birth Weight (3-year running average)

Figure 3c. Dublin, 1869-1930: Percentage Stillborn (3-year running average)

Figure 4c. Boston, 1872-1900: Percentage Stillborn (3-year running average)

Records of the Philadelphia Almshouse Hospital over the period 1848 to 1873 were examined by Goldin and Margo (1989). They report an average weight of 3,400 grams (all weights, including those under 1,500 grams), 10 per cent of which fell into the low birth weight category. No statistically significant time trend was found in their study.
About 56 per cent of their sample was unmarried, one-half was foreign born, and two-thirds were under age 25.

Figure 5a. Montreal, 1851-1900: Mean Birth Weight

Figure 5b. Montreal, 1851-1900: Percentage Low Birth Weight (3-year running average)

Figure 5c. Montreal, 1851-1900: Percentage Stillborn (3-year running average)


Although precise comparisons of the quality of life across long spans of time are impossible, it is reasonable to suggest that the living standards among the poor in the industrializing countries of the nineteenth century may have approximated that of typical individuals in poor developing countries of the late twentieth century. In both situations, incomes were so low that a large majority of income was spent on food, work was often physically arduous, and personal hygiene and public health practices were meager by modern standards. Thus, it is not surprising to find many instances in the past in which average birth weight fell below 3,400 grams, the incidence of low birth weight exceeded 5 per cent, and the share of stillborns was more than 2 per cent.

Although the number of cities studied is rather small, some patterns across cities are worth noting. The time patterns differed, with the large declines registered in Edinburgh and especially Montreal. Nowhere did outcomes improve substantially, but a slight upward trend was detectable in Dublin. Newborns were on average heaviest in the United States and lightest in continental Europe, and the largest difference amounted to nearly 350 grams between Boston and Vienna. The incidence of low birth weight ranged from approximately 6 per cent in Boston and Montreal to approximately 13 per cent in Vienna. Stillbirths were most frequent in Edinburgh (about 12 per cent) and least likely in Boston (about 5 per cent). However, the per cent stillborn climbed to about 18 per cent in Edinburgh between 1910 and 1915 and were as low as 2 per cent among patients at the New England Hospital in the late 1870s.

Fortunately, the maternity hospitals usually recorded the mother's age; often characteristics of the child, such as parity and gender; and sometimes features of the mother such as occupation, religion, ethnic background, or marital status. This additional information allows one to measure some influences on birth outcomes using multiple regression analysis. Male children were in general heavier, having an advantage of 101 grams (Montreal) to 131 grams (Vienna). Higher parities were also 13 grams (Edinburgh) to 56 grams (Boston) heavier. Within the age range studied, older women tended to have slightly heavier babies (3.8 to 6.2 grams per year of age). However, the Philadelphia study reports a nonlinear relationship, with maximum weight occurring at 31 years of age (it cannot be determined from the information available whether this relationship was statistically significant). Mothers who were ill at the time of delivery had smaller babies, ranging from 124 grams in Boston to 433 grams in Dublin. In Philadelphia, mothers ill with a venereal disease had births weighing 640 grams less than those without a venereal disease reported; alcoholics had births that were 1,174 grams lighter than non-alcoholics; and twins weighed 874 grams less than singleton births. Occupational background also influenced birth outcomes. Mothers who were domestic servants had heavier (111 grams) babies in Edinburgh, perhaps because they often had good access to food. In Vienna, mothers with food handling occupations had babies that were 68 grams heavier.

Concluding remarks

Historical documents also provide numerous indirect sources of information about intrauterine growth. Because conditions that influence growth prior to birth often persisted after birth, measures such as birth weight may be highly, positively correlated with growth after birth. If true, then early childhood heights and even adult height-when combined with other sources of information-may provide some insights into growth prior to birth. For example, this line of reasoning suggests that a majority of births to American slave women were of low birth weight (Steckel, 1986, 1987). The slave women experienced highly seasonal patterns of work, diet, and disease that probably created intense stress on intrauterine growth during certain periods of the year, particularly late winter to early spring, and late summer. Numerous historical studies of growth in childhood are discussed in Steckel (1987), and various papers in Steckel and Floud (1997) give international perspective to long-term trends in adult heights.


Goldin C and Margo RA (1989): The poor at birth: Birth weights and infant mortality at Philadelphia's almshouse hospital, 1848-1873 Explorations in Economic History 26, 360-379.

Steckel RH (1986): Birthweights and infant mortality among American slaves. Explorations in Economic History 23, 173-198.

Steckel RH (1987): Growth depression and recovery: the remarkable case of American slaves. Annals of Human Biology 14, 111-132.

Steckel RH and Floud R (eds) (1997): Health and Welfare during Industrialization. University of Chicago Press: Chicago.

Tanner JM (1981): A History of the Study of Human Growth. Cambridge University Press: Cambridge.

Ward W Peter (1993): Birth Weight and Economic Growth. University of Chicago Press: Chicago.


Whereas the desire to assess the chances of survival of newborns and infants can be traced back to Antiquity, the systematic measurement and weighing of newborns is a much more recent practice, and the methods used initially were not standardized. There are some records of birthweight from the 18th century, but a variety of different weight units were used at that time and their comparison and translation into contemporary weights is not easy. Measurements of children's length are more recent still, and early ones are extremely low, suggesting that newborns and infants were probably not stretched out before they were measured.

As Steckel shows, clinical records in some hospitals contain birthweight data in the latter half of the 19th century, but these were not used for clinical purposes. Many of these records were unfortunately thrown away or destroyed. The use of anthropometric data of newborns and infants as prognostic indicators began only after World War II, perhaps because, in industrialized countries, they then became fairly good predictors of survival into adult life. They also started to be used as outcome variables in the assessment of the efficacy of interventions. In some traditional societies people, for various reasons, are still reluctant to have their babies weighed and measured. This may change as more and more babies are delivered in hospitals.

High levels of physical activity and standing on ones feet during much of the day is more frequent in pregnant women who are subjected to other unfavorable circumstances as well. Their independent effect on fetal growth is therefore difficult to assess. Heavy work in pregnant women in developing countries appears to have an effect on birthweight less on its own than through the periods of negative energy balance it frequently entails. Reduced birthweight could, however, also be shown in well-off female athletes with ad libitum access to food, but in this case thinness was a confounding factor and low birthweight was primarily due to prematurity.

To some extent at least, the interests of the mother (low risk of cephalo-pelvic disproportion and other complications in labor) and of the child (to be born with a high birthweight) are in contradiction. Traditional views tend to favor a reduction of dietary intake towards the end of pregnancy (e.g. by 100 to 150 kcal/d in rural Guatemala) and thus to benefit the health of the mother. This tradition that seems to have originated in India around 1500 BC is still widespread and also had its proponents in Western countries earlier in this century. In Southeast Asia it becomes confounded with the 'hot vs cold food' concept and the belief that pregnant women (contrary to lactating women) should eat 'cold' foods that have a low energy and protein content. Sometimes mothers admit that they reduce their food intake to reduce the risk of complications in labor, others argue that they simply find it difficult to eat more of the bulky staple food with low energy density when much of the abdominal space is taken up by the baby towards the end of pregnancy. Studies of dietary supplementation of marginally nourished pregnant women showed that birthweight can be substantially increased without a simultaneous increase of the risk of cephalo-pelvic disproportion in labor.

Where it is possible to distinguish between IUGR and prematurity it is clear that in industrialized countries the left-hand tail of the birthweight distribution is primarily due to prematurity, which has shown only small changes attributable to interventions and secular changes. In both developed and developing countries prematurity increases early mortality more than growth retardation.

Effects of unfavorable circumstances (e.g. smoking and unfavorable psychosocial conditions) and favorable interventions (e.g. food and Zn supplementation) are stronger and often only detectable in women with low BMI at conception. Recommending high pregnancy weight gains in women with a high BMI at conception is unlikely to affect birthweight, but puts the mother at a high risk of retaining considerable amounts of fat post partum. Current dietary recommendations suggest that European scientists are more sensitive to this possible negative side effect than their colleagues in the US.

Contents - Previous - Next