Urban Infant Mortality in Imperial Germany

© 1994 The Soarry for the Social History of Midinnt

Urban Infant Mortality in Imperial Germany

By JORG P. VOGELE*

SUMMARY Infant mortality in Imperial Germany started to decline in urban areasfrom the 1870s onwards, whereas national rates did not decrease before the beginningof the twentieth century Therefore, key explanatory factors determining the levels andtrends of infant mortality are investigated in an urban context These include the declineof birth rates, the legitimacy status of infants, feeding practices, environmental con-ditions, and economic growth. Through a rising living standard and by creating ahealth-preserving environment, urban populations lost their traditional disadvantagein survival chances. This went so far that even high risk factors, such as the abandon-ment of breastfeeding, could be counterbalanced. In this sense, a study of past urbanhealth conditions functions as a paradigm for the situation in industrialized societies

KEYWORDS infant mortality, Germany, urban history, nineteenth century.

Infant mortality is commonly seen as one of the most sensitive indicators ofsocioeconomic and environmental conditions.1 High infant mortality ratesduring the nineteenth century shed an informative light on contemporaryliving conditions. In some years only one third of all infants born survivedthe first year and often only one half of those born in a particular year reachedmaturity. Traditionally, the population of towns was a high risk group andadult men and infants were especially at risk in the urban environment.2

Industrialization and urbanization radically changed urban living conditions.A rapidly increasing population density in the course of urbanization wasaccompanied by an accelerated transmission of disease. Yet, urban agglomera-tions of Western Europe possessed the financial potential and the innovativepower to carry out adequate measures to secure health on a large scale.Therefore, the study of urban development may function as a paradigm forthe general situation in industrialized societies.

In the period under investigation, from 1870 to the First World War, infantmortality in Germany declined substantially from 21 deaths per 100 births in1875 (in Prussia) to 15 deaths per 100 in the German Empire in 1913 Keyexplanatory factors, cited in recent studies, as determining these levels andtrends of infant mortality include, for example, legitimacy status of infants,fertility, public health services, feeding practices, parental wealth and occupa-

* Heinnch-Heine-Umversitat Dusseldorf, Medizinische Einnchtungen, Institut fur Geschichteder Medizin, Postfach 101007, D-40001 Dusseldorf, Germany

1 This research was made possible by The Wellcome Trust, London, and the Alexander vonHumboldt-Stiftung, Bad Godesberg I am grateful to W R Lee and Fiona Lewis, University ofLiverpool, for general and linguistic advice I would also like to thank the referees for theirvaluable comments and suggestions

2 J Vogele, 'Die Entwicklung der (groB)stadtischen Gesundheitsverhaltmsse in der Epoche desDemographischen und Epidemiologischen Ubergangs', in J Reulecke and A Castell, eds , Stadtund Gesundheit Zum Wandel von " Volksgesundheit" und kommunaler Gesundheuspolitik tm 19 undfruhen 20 Jahrhundert (Stuttgart, 1991), 21-36

0951-631X Social History of Medicine 07/03/401-425

402 Jorg P. Vogele

tion, and general attitudes towards life and death.3 Social and regional varia-tions, as well as the course of infant mortality in Germany, have already beendiscussed in great detail. Imhof attributes regional differences in infant mortalityto a large extent to different attitudes towards life and death.4 The course ofinfant mortality in Germany has been investigated by Spree and Stockel. Spreeuses infant mortality as an indicator to analyse social inequality,5 and in hismost recent contribution stresses that regional disparities were diminishing inPrussia during the late nineteenth century, whereas social inequality increased.6

Stockel discusses various factors influencing infant mortality in Berlin as a casestudy, and employs a wide range of quantitative and qualitative sourcematerial.7 The most advanced methodological approach has been provided byKintner, who has developed a model concerning the determinants of infantmortality in Imperial Germany The empirical evaluation using multiple re-gression analysis, however, remains unsatisfactory. Such an analysis requiresa set of well-refined indicators over time and space, and the existing sourcematerial does not fulfil such a prerequisite. The following analysis thereforeuses a classical approach, discussing the various indicators separately. It has tobe emphasized in this context that all the factors under discussion had aplausible effect on the level and trends of infant mortality and that their relativeimportance changed over time and space.

Existing studies have usually collected their data on the level of largeradministrative units, the Regierungsbezirke, which comprised various townsand cities as well as rural areas. They were therefore heterogeneous units.However, it is important to emphasize that the decline in infant mortalitystarted earlier and was more pronounced in towns than in rural areas, withthe consequence that the traditionally high urban infant mortality rates disap-peared. After the turn of the century the towns showed increasingly lowerrates than rural areas or indeed, the national aggregate. In this respect thetowns and especially the large cities (Grofstadte), traditionally regarded asbeing particularly unhealthy, operated as a role model in the dramatic change

3 For a comprehensive account see H J Kintner, 'The Determinants of Infant Mortality inGermany from 1871 to 1933', unpublished Ph D thesis, University of Michigan (1982) For theinternational discussion see R Woods, P A Watterson and J H Woodward, 'The Causes ofRapid Infant Mortality Decline in England and Wales', Population Studies, 42 (1988), 343-66, 43(1989), 113-32

4 A E. Imhof, 'Unterschiedliche Saughngssterbhchkeit in Deutschland, 18 bis 20 Jahrhundert-Warum 5 ' Zeuschnft fur Bevolkerungswissenschaft, 7, 3 (1981), 343-82 See also U Ottmuller,Spetkmder-Gedeihkinder Kommuiukattotislheoretische Uberlegungen zu Gestalt und Funktwn fruhkind-licher Sozialisation im bauerluhen Lebenszusammenhang des deutschsprachigen 19 undfruhen 20 Jahrhun-derts, Ph D thesis, University of Berlin (1986)

5 R Spree, Health and Social Class m Imperial Germany A Social Study of Mortality, Morbidityand Inequality (Oxford, 1988) R Spree, 'Die Entwicklung der differentiellen Saughngssterbhch-keit in Deutschland sen der Mitte des 19 Jahrhunderts (Ein Versuch zur Mentahtatsgeschichte),in A E Imhof, Mensch und Gesundhett m der Gcschichte (Husum, 1980), 251-78

6 R Spree, 'On Infant Mortality Change in Germany since the Early 19th Century', unpub-lished paper (Munich, 1993)

7 S Stockel, 'Saughngssterbhchkeit in Berlin von 1870 bis zum Vorabend des ersten Weltknegs— Eine Kurve mit hohem Maximum und starkem Gefalle', Berlin-Forschungen, 1 (1986), 219-64

Urban Infant Mortality in Imperial Germany 403

towards modern health conditions. For example, the new industrial areas ofthe Rhineland and Westphalia were leading the way in respect of both theorganization and the actual implementation of health-related infrastructuralmeasures.8 At the same time these towns offered a relatively high standard ofliving. In Diisseldorf, for example, for many years the average wages for daylabour even surpassed those paid in Berlin. It is for this reason that themechanisms behind this development have to be investigated in the contextof specific urban-industrial living conditions and urban measures undertakento secure or improve health conditions. In order to evaluate their impact thefollowing analysis will discuss a broad range of variables affecting infantmortality within an urban context. Therefore, the paper will first describe theinfluence of socio-demographic factors, namely (1) the decline of the birthrate, which started earlier in urban areas, and (2) the legitimacy status ofinfants. The focus will then shift to (3) comparative feeding practices, (4)improved environmental conditions as a result of sanitary reforms, and (5) therise of urban living standards during industrialization. These potential deter-minants of the registered decline in infant mortality will be analysed in thefollowing discussion within the framework of a historical epidemiology.

Specific developments in the ten most populous towns (in 1910) as well asaverage conditions in all towns with a population exceeding 15,000 habitants,will form the basis of the following analysis. A direct comparison with ruralareas, however, will be provided by other studies. Infant mortality in Germanyfollowed a strong regional pattern with substantial variations over short dis-tances. Therefore a reasonable urban-rural comparison could only be obtainedby investigating a specific town and its surrounding rural areas, using data onthe level of larger registration areas. For the largest towns, however, this isnot an adequate procedure as the areas directly surrounding often had an urbancharacter as well. Furthermore, as the conditions in the largest towns func-tioned as a role model,9 the traditionally applied urban-rural dichotomywould underestimate their primary role within the context of the secularmortality decline. Therefore, the conditions in Prussia and in the GermanEmpire will form the comparative framework, providing trends rather thanexact statistics as Prussia and Germany were becoming increasingly urbanizedduring the period under investigation.

In order to calculate infant mortality rates, the number of infants under oneyear of age who died within one specific year will be set into relation to thenumber of births within this year. Still-births were generally excluded fromcontemporary calculations. Their registration, however, differed regionallyWhereas in Catholic areas the number of registered still-births was reduced bythe accepted custom of emergency baptism, infant mortality was generally

8 A Labisch, 'Kommunale Gesundheitssicherung lm rheimsch-westfahschen Industnegebiet(1869-1934) —ein Beispiel zur Soziogenese offenthcher Gesundheitsleistungen', in H Schade-waldt and K -H Leven, eds , XXX lnternationaler Kongrefi fir Geschtchte der Medizw, Dusseldorf31 8-5 9 1986, Actes/Proceedings (Dusseldorf, 1988), 1077-94

9 See Jorg Vogele, Urban Mortality Change m Britain and Germany, 1870-1910 (Liverpool,forthcoming)

404 Jorg P Vogele

underrated in regions where all deaths that occurred within a mandatory three-day registration period were treated as still-births according to the Civil Code.10

Data was compiled from the publications of the Kaiserhche Gesundheitsamt(Imperial Health Office), which published the number of births and deaths forall towns exceeding 15,000 inhabitants on an annual basis since 1877. Thecauses of death were collated from official Prussian statistics, as well as fromthe statistical yearbooks of individual towns.

Infant Mortality Change

Infant mortality in Germany generally rose after the mid-nineteenth centuryand reached its pinnacle in the 1860s and 1870s (Table 1).u The rates generallyfluctuated between 10 and 30 deaths per 100 births. Especially high rates wereregistered by the states in the south-east and the east of the German Empire.Around the mid-nineteenth century peak infant mortality rates (per 100 births)were recorded in Wurttemberg with 35.4 (1858-66), in Bavaria with 30.7(1827-69), in Baden with 26.3 (1856-63), and in Saxony with 26.3 (1859-65).At the other end of the scale were Oldenburg with 12.3 (1855-64), Schleswig-Holstein and Lauenburg with 12.4 (1855-59). Prussia with 20.4 per 100 (1859-64) occupied an intermediate position.12 The eastern provinces of Prussia,however, registered very high infant mortality with maximum rates between24 and 27 recorded in Silesia after the middle of the century, in contrast tosignificantly lower rates in the west, which fluctuated in the Rhineland, forexample, between 13 and 20 per 100.13 Due to these high infant mortalitylevels in certain regions, the German Empire showed up badly in an inter-national context, especially when compared to Scandinavia or England.

The situation in towns was regarded as particularly hazardous. The levelsof infant mortality, however, differed widely between towns, and may havebeen to a large degree the result of regional factors rather than the reflectionof specific local conditions in particular towns.14 In areas with high infantmortality, the rates were lower in an urban environment, in low infant mortalityregions the opposite was the case. In Bavaria, Wurttemberg and Baden, i.e.the non-breastfeeding areas, infant mortality was consistently lower in urbancommunities when compared to rural areas 15 In the three last decades of the

10 G Mayr, 'Die Sterbhchkeit der Kinder wahrend des ersten Lebensjahres in Siiddeucschland,insbesondere in Bayern', Zettschnft des komglich bayenschett stattstischen Bureau, 2 (1870), 201-47,p 203. For frequencies and causes of late nineteenth-century still-births see F Pnnzing, 'DieUrsachen der Totgeburt', Allgemewes Stattsttsches Archw, 7 (1914), 21-49

11 F Pnnzing, 'Die Entwicklung der Kindersterbhchkeit in den europaischen Staaten', Jahr-bucher fur Nattonaloleotiomie und Statisttk, III ser , 17 (1899), 577-635, pp 583-4, W R Lee,'Germany', in W R Lee, ed , European Demography and Economic Growth (London, 1979), 144-94, p 155

12 Mayr, 'Sterbhchkeit', pp 207-813 Pnnzing, 'Entwicklung', p 587 For a systematic analysis of regional and social differentials

in Prussia see Spree, 'On Infant Mortality'14 J Knodel, 'Town and Country in Nineteenth-century Germany A Review of Urban-Rural

Differences in Demographic Behaviour', Sofia/ Science History, 1, 3 (1977), 356-82, p 37715 Mayr, 'Sterbhchkeit', p 226, K Maier, 'Die Sterbhchkeit der Kinder lm ersten Lebensjahr


TABLE 1 Infant mortality rates w Germany (per 100 births)

Prussian provincesEast PrussiaWest Prussia

City of BerlinBrandenburg (wo Berlin)PomeramaPosenSilesiaSaxonySchleswig-HolstemHannoverWestphaliaHessen-NassauRhinelandHohenzoliern

1811-20

Prussia 17Bavaria -Wiirttemberg 32BadenSaxony -

1821-30

17§29

23-

1718221614182318_-

14-

15-

1819-33

1831-40

1830_-

27

1834-48

1919231715192418

*13_

14-

15-

1841-50

193035-

26

1849-63

2122251916222420

**12+ 13

14-

16-

1851-60

2031-

2526

1864-70

2324301919242621

+ + 15+ + 16

16+ + 18

1834

1861-70

2133362827

1871-80

2131322629

1871-75

2224342520242723151718182033

1881-90

2128272328

1876-80

2223302320222621151415161732

1891-1900

2027242227

1881-88

22232725202226221515151518-

1901-5

1924222025

1889-95

2223252522212522161515141825

1906-10

1722181720

* 1845-54 **1855-59t 1853-55 ft1867-70§ 1827/28-33/34

Source* F Pnnzing, "Die Entwicklung der Kindersterbhchkeit in den europaischen Staaten', Jahrbucher furNatioitahkotwmte und Statnttk, III ser , 17 (1899), 577-635 , p 587, F Pnnz ing , Handbuch der medizuuschenStatntik (Jena, 1931) 2nd edn , p 375, G Mayr, 'Die Sterblichkeit der Kinder wahrend des ersten Lebensjahresin Suddeutschland, insbesondere in Bayern', Zeihclmfi det komglieh bayemcheii lUtutuchen Bureau, 2 (1870), 201-47, pp 208, 238, F Burgdorfer, 'Geburtenhaufigkeit und Sauglingssterblichkeit mit besonderer Berucksich-tigung der bayenschen Verhaltnisse', Allgemeuiei Stati>thchei Arduv, 7 (1914), 63-154, p 103

nineteenth century urban-rural differences in infant mortality were predomi-nantly marked in towns by a higher risk of dying after the initial month (post-neonatal mortality).16 Still-births and neonatal mortality were consistentlyhigher in rural areas, yet urban-rural differences were much less pronounced.17

Taking all German towns over 15,000 inhabitants, infant mortality declinedin the last decades of the nineteenth century, and this trend accelerated afterthe turn of the century. In contrast, the decline for the whole of Prussia did

in Bayern', Journal for Kinderkrankheiten, 57 (1871), 153-98, p 181, A Wiirzburg, 'Die Sauglings-sterblichkeit im Deutschen Reiche wahrend der Jahre 1875 bis 1877', Arbeiten aus dent KaiserlichenCesmidheitsamte, 2 (1887), 208-22, 343-6, 4 (1888), 28-108, p 363, F Pnnzing, 'Die Kinderster-bhchkeit in Stadt und Land', Jahrbucher fur Natwualokoitonue und Slatistik, III ser , 20 (1900), 593-644, p 633

16 Preufiische Slatistik, 188 (1904), p 126, Pnnzing, 'Kindersterbhchkeit', pp 610-1217 H Bleicher, Statistische Beschreibung der Stadt Frankfort am Mam und threr Bevolkerung (=

Beitrage zur Statistik der Stadt Frankfurt am Main, Vol 1, Part II) (Frankfurt a M., 1895), pp476—7, A. Schlossman, 'Studien uber Sauglingssterblichkeit', Zeitschnft for Hygiene und Infektwns-kranleheiten, 24 (1897), 93-188, pp 126-9, Pnnzing, 'Kindersterbhchkeit', pp 602-9, Knodel,'Town and Country', p 373

406 Jbrg P. Vogele

Death rate (per 100 births)

" Towns > 15,000 inh °' Prussia/German Emp

FIG 1 Infant mortality in German towns and in Prussia/German Empire 1877-1913 (Prussia1877-1900, German Empire 1901-1913)

not begin before the first decades of the twentieth century (Figure I).1 8 In thecourse of the twentieth century the situation changed and infant mortahty ratesin towns dropped below those in rural areas:19 for example, in Berlin infantmortality was 90.1 in 1924-26 (per 1,000 births), in comparison to 103.8 inPrussia as a whole.20 Even the death rate of illegitimate infants — with tra-ditionally very high rates — declined in an urban environment, whereas itincreased in the countryside 21

In order to gain some insight into the mechanisms behind these changes,the causes of death among infants have to be taken into account Such ananalysis, however, is confronted with two problems, namely (1) the officialcertification of death, and (2) the causes-of-death classification itself.

1) Apart from Prussia, official death certification was obligatory in the largerGerman states. However, it was only carried out by a doctor if the person hadbeen in medical treatment. As a doctor was rarely consulted in the case of asick infant,22 it is most likely that most causes of death among infants werenot certified by an expert, but often registered according to information

18 Sources Veroffentlichungeii des Kaiserluhen Gesutidheitsamtes (Balaget:), (1878-1914), FRothenbacher, 'Zur Entwicklung der Gesundheitsverhaltmsse in Deutschland seit der Industnal-lsierung', in E Wiegand and W Zapf. eds . IVandel der Lebensbedmgungen in Deutschland Wohl-fahrtsentuncklung sett der Industnahsierung (Frankfurt/M , 1982), 335—124, p 396

19 R Kuczynski, Zug nach der Stadt Statistische Stitdien uber Vorgange der BevolkenmgsbewegungiMI Deutschen Reiche (Stut tgar t . 1897). pp 198-214, Vogele , 'En tw ick lung ' , p 26

20 F P n n z i n g . Handbuch der medizimschen Statistik (Jena, 1931) 2nd edn , p 56921 P n n z i n g , 'Kinders terbhchkei t ' , p 613 and Table 322 Schlossmann, 'Studien', p 178

Urban Infant Mortality m Imperial Germany 407

proffered by parents or relatives.23 In Lower Bavaria in 1866-67 a doctor wasonly consulted in 3 per cent of all cases of infant sickness.24 In Wiirttembergonly 38 per cent of all infants dying in 1900 received medical treatment, incomparison to 70 to 90 per cent in the higher age-groups. 2D In Dresden in1891-96 death was not certified by a doctor in 49 per cent of all infant deaths,whereas in the higher age-groups this was the case in only 2 per cent of alldeaths.26 The situation was especially unsatisfactory in rural areas with a lowerdoctor-patient ratio.27

2) Causes of death were registered in a terminology far removed frommodern diagnosis. The use of a completely different classification system,often describing final symptoms (convulsions) or the period in life when deathset in ('teeth' as a cause of death does not refer to an actual cause of death, butrather indicates that death occurred while the infant was getting his teeth),28

rather than the actual sense of death in a modern sense, makes translation intopresent-day terminology at times impossible Yet, it has to be stated that incontrast to the 999 official causes of death in modern death registration, wehave only to deal with a relatively small number of causes for the period underconsideration. The most important of these diseases were the classic acute andchronic infections, so that a diagnosis was definitely not as taxing as it mightbe today, with the predominance of cardio-vascular disease and cancer. Acareful analysis of the historical cause-of-death data is therefore both feasibleand legitimate, and provides valuable information

The predominant causes of death among infants were gastro-intestinal diseasesand weakness of life (Lebenschwache),29 often linked with Abzehnmg (emaci-ation, atrophy), which most likely was a result of an illness of the digestivesystem (Table 2) 30 In Upper Bavaria and Swabia around the middle of thenineteenth century (1867/69), for instance, one quarter of all infants died ofdiarrhoea, atrophy and Fraisen (an associated condition). Thus, digestivediseases comprised up to 70 per cent of the whole disease panorama.31 Withsubstantially lower rates, diseases of the respiratory system followed in second

23 Pnnz ing , 'Kindersterbl ichkei t ' , p 634, F P n n z i n g , Handbuch der mediziinscheti Statistik (Jena,1906) 1st edn , pp 3 2 3 - 5

24 Maier , 'S te rbhchken ' , p 19325 F Prmzing , Handbuch (1906), p 32226 Schlossmann, 'S tudien ' , p 17827 W u r z b u r g , 'Saughngss terbhchkei t ' , p 49, Pnnz ing , Handbuch (1906), p 32328 Schlossmann, 'S tudien ' , p 17929 C o n t e m p o r a r y experts considered convuls ions , emaciat ion, a t rophy, and teeth as causes of

death fol lowing sickness of digestive disease, and consequent ly subsumed those diseases underthe digestive disease g r o u p See Prmzing . 'En tw ick lung ' , 577-635 H J Kintner confirmed thisview applying regression analysis on cause-specific infant morta l i ty rates for var ious Ge rmanadminis t ra t ive areas See H J Kintner , 'Classifying Causes o f Death dur ing the Late Nineteenthand Early Twen t i e th Centur ies T h e Case of G e r m a n Infant Morta l i ty ' , Historical Methods, 19, 2(1986), 4 5 - 5 4

30 C Flugge, 'Die Aufgaben und Leistungen der Mi lchs tenhs ierung gegenuber den D a r m -krankheiten der Saughnge', Zettschrtft jiir Hygiene und Infectionskrankheiten, 17 (1894), 272-342, p275, Wurzburg, 'Saughngssterbhchkeit', pp 48-52, Pnnzing, 'Kindersterblichkeit', pp 636-7

31 Mayr, 'Sterbhchken', p 218

408 Jorgi P. Vogele

TABLE 2 Causes of death among infants (per 10,000 births)

Cause of death

Weakness of lifeSmallpoxScarlet feverMeasles/RubellaDiphthena/CroupWhooping coughTyphoid feverDysenteryDigestive diseases*ScrofulaTuberculosisBrain and nervesRespiratory

diseases+Influenza

1877

60716

1720392

141257

82996

2153

Other communicablediseases

ViolenceOther causes

All causes

06

183

2502

Towns

1885

70114

302531

13

1156103077

24023

23

166

2422

1900

60902

329

3300

1114214484

2700

76

133

2362

1907/10

37301

248

3600

6200

3876

2343

3111

214

1668

1877

3990

253381934

10972

61976

64-

-3

216

2000

in ten large

Prussia

1885

4411

1939797626

10008

2066

96-

-3

184

2040

1900

4610

10192871

12

1126131957

163-

-3

150

2125

German

1907

353-6

1911450-

405-

2342

1893

74

576

1683

towns and in Prussia

DifferenceTowns-Prussia

1877

2081

-19-16-61-54- 2

4285

21020

151_

-3

-44

502

1885

2600

-15- 9

-54-45- 1- 3156

21011

144-

-0

-23

382

1900

1480

- 813

-19-38- 1_2

-128

2527

107_

-

-23

237

1907

20-

—55

- 3- 9

0-

215-

1534

450

247

-367

-15

* including cholera nostras, diarrhoea, convulsions (Prussia 1907 excluding convulsions)+ including bronchitis, catarrh of the lungs, pneumonia, pleurisy, other diseases of the lungs

Sources. Preufjischc Statistik, Mittheilungen des Statishschen Bureaus der Stadt Dresden, Jahresbericht desstattstischen Amtes der Stadt Dresden, Statistisches Jahrbuch fur die Stadt Dresden, Statistik des Hamburgt-schen Staates, Statistik des Deutschen Reuhs, Mittheilungen des statistischen Bureaus der Stadt Leipzig,Statistisches Jahrbuch der Stadt Leipzig, Mitteilungen des stat Bureaus der Stadt Munchen, Mittetlungendes stat Amtes der Stadt Munchen

place. In third place were diseases of the brain and nerves.32 Whereas inPrussia, when compared to the urban sample, mortality from the classicinfectious diseases of childhood was clearly higher (scarlet fever, measles andrubella, diphtheria and croup, whooping cough), it was especially the gastro-intestinal disease complex that was responsible for the higher urban overallmortality, followed by weakness, and diseases of the respiratory system (Table2). The frequency of digestive diseases increased with urban size: in Prussia in1880, 91.6 out of 1,000 live births died of diarrhoea and dysentery in townswith a population exceeding 100,000 inhabitants, 45.8 in towns between20,000 and 100,000 inhabitants, 19.1 in towns below 20,000 inhabitants, and8.3 in rural communities.33 Until 1900 the reduction of high infant mortalitylevels in urban areas was principally brought about by a decrease in gastro-intestinal diseases, although the massive decline of this cause of death occurred

32 Diseases of the nervous system might also, at least partially, have included intestinal diseasesSee Kmtner, 'Historical Methods', p 46

33 Wurzburg, 'Saughngssterbhchkeit', p 56.


Coefficient of variation

0.75

0,5

1895

Year

•"*• Crude death rate "T Infant mortality ~*~ Tuberculosis —Digestive diseases

FIG 2. Urban variations in mortality in German towns 1877-1913

after the turn of the century. This is also evident in Prussia, although thepronounced increase in mortality from 'other causes' may be the result ofchanges in cause-of-death registration (referring especially to convulsions),34

so that a direct comparison is difficult, if not impossible.The digestive diseases were also mainly responsible for the wide inter-city

variations. Expressed as a coefficient of variation, i.e. standard deviation ofthe variable divided by the mean, Figure 2 demonstrates this using a sampleof 19 towns with over 15,000 inhabitants.35 Variations in mortality fromdigestive diseases (for all age-groups) and infant mortality declined from thebeginning of the 1890s to a lower level, whereas differences in mortality fromtuberculosis (for all age-groups) between the towns even increased. With thisin mind, attention must be focused on changes in the urban environment,particularly during the late-nineteenth century, because of their potential im-pact on the decrease in digestive diseases. In this context, the analysis will nowfocus on the influence of the birth rate on infant mortality, the legitimacystatus of infants, the role of feeding practices, as well as the affects of sanitaryreforms and of rising living standards on mortality change.

34 'Preufien ErlaB, bet r die Neubea rbe i tung des Verzeichmsses der K r a n k h e n e n und T o d e s u r -sachen', Veroffenthchungeti des Kaiserlichen Gesundhettsamtes, 28 (1904), 645-51, Preufiische Statisttk189 (1905), p 6 See also Kintner , 'Historical M e t h o d s ' , p 47

35 Source Veroffentlichutigeii des Kaiserhchen Cesundheitsamles (Beilagen), (1878-1914)

410 Jorg P. Vogele

Determinants of Infant Mortality Change

Infant mortality and birth rates

The modern secular fertility decline in Germany started in urban areas, andespecially in the larger towns. The specific interrelationship between birthrates and infant mortality, however, is still far from clear, both with respectto the degree of relationship and the direction of the influences.36 A declinein infant mortality may lead to more extended birth intervals, either by a moreor less conscious limitation of reproduction behaviour, through birth control,or by lactation-amenorrhoea during the period of breastfeeding. In this sense,intensive breastfeeding reduces not only infant mortality, but also fertility.37

A smaller number of children again might have a positive influence on thesurvival chances of infants, in that greater spacing between births might leadto more intensive care of already existing children.

For Germany at the end of the nineteenth century, a strong statisticalassociation between the birth rate and infant mortality can be established insome areas.38 In Bavaria during the period 1871 to 1910, for example, thecoefficient of correlation between the birth rate and infant mortality was upto 0.7777 (sig=0.000, n=40),39 i.e. a high birth rate coincided with high infantmortality, or as expressed by a contemporary commentator, 'mass birthsprovoke mass deaths of infants, and these again provoke mass births'.40 In theurban milieu there was no direct association.41 A reduced birth rate did notobviously lead to a decline in infant mortality. Nor is there any direct verifica-tion of the plausible argument that a decline in infant mortality brought aboutlower birth rates. A correlation between the birth rate and infant mortality innineteen German towns over 15,000 inhabitants (among them the ten largest)suggests a figure of only 0.3657 between 1877 and 1913 (sig. =0.000; n=691).42

Using fertility rates for the ten largest towns in selected years, the analysisprovides a similarly low coefficient.43 This is at least partly the result ofregional particularities. In the Rhineland high birth rates were accompaniedby low infant mortality rates 44 the towns of this region, among them, Diissel-dorf, registered the highest birth rate in the sample, far above the average for all

36 J Knode l , The Decline of Fertility in Germany, 1871-1939 (Pr ince ton , 1974), pp 1 4 8 - 8 7 ,Knodel, 'Town and Country', pp 344—51, Woods, Watterson and Woodward, 'Causes', pp 121-6

37 J . K n o d e l , Demographic Behavior in the Past A Study of Fourteen German Village Populations inthe Eighteenth and Nineteenth Centuries ( C a m b r i d g e , 1988), p p 393—405

38 Mayr, 'Sterbhchkeit', p 23039 Source F Burgdorfer, 'Geburtenhaufigkeit und Saughngssterbhchkeit mit besonderer

Berucksichtigung der bayenschen Verhaltmsse, Allgememes Statistisches Archiv, 7 (1914), 63-154,pp 102-3

40 G Mayr quoted after F Burgdorfer, 'Geburtenhaufigkeit und Saughngssterbhchkeit mitbesonderer Berucksichtigung der bayenschen Verhaltmsse', Allgememes Statistiches Archiv, 7(1914), 63-154, p 97

41 C Ba l lod , Die mittlere Lebensdauer m Stadt und Land (Le ipz ig , 1899), p p 4 1 - 242 M i s s i n g va lues = 12 Source Veroffentlichungen des Kaiserhchen Gesundheitsamtes (Beilagen),

(1878-1914)43 R = 0 3694, sig (two-tailed) = 0 045 Selected years and sources see Table 244 Pnnzing, 'Kindersterbhchkeit', p 640



40

1875 1880 1885 1890 1895 1900 1905 1910 1915

"""Hamburg —Frankfurt Dusseldorf "•" Berlin

FIG 3 Infant mortality in selected German towns 1877-1913

towns over 15,000 inhabitants, together with low infant mortality rates.Furthermore, they were frontrunners in the general overall decline in infantmortality after the turn of the century (Figure 3) (see fn. 42)

Legitimacy status of infants

Illegitimate infants traditionally had an especially slim chance of survival, asthe lack of care went along with a disadvantaged economic situation. Corre-spondingly, the mortality rates of illegitimate infants significantly exceededthose of the legitimate (Table 3). In Prussia the respective death rate forlegitimate and illegitimate infants was as follows: 19.4 and 35.3 (per 100 births)in 1875-80, 19.4 and 35.4 in 1881-90, 19.0 and 35.5 in 1891-1900; and finally16.8 and 30.8 in 1901-10 45 As the proportion of illegitimate infants remainedfairly stable in the period under investigation,46 its potential impact has to betaken into account merely in regional or local comparisons.

Again there were vast local and regional variations, with a traditionally highproportion of illegitimate infants in the eastern and south-eastern parts ofGermany and a low proportion in the western parts. In areas with low overallillegitimacy, the level was higher in the urban than in the rural sector; other-wise, the reverse was true.47 For example, the Regierungsbezirk Dusseldorf

45 S Engel and H Behrendt , 'Saughngsfursorge ' , in A Gottstein, A Schlossmann and LTeleky, eds , Handbuch der Sozialen Hygiene, Vol 1 (Berlin, 1927), 28-194 , p 40

46 F Prinzing, 'Ehehche und unehehche Fruchtbarkei t und Aufwuchsziffern in Stadt und Landin PreuBen', Deutsche mediziiusche Wochenschnft, 13 (1918), 3 5 1 - 4 , p 353

47 Knodel, 'Town and Country', p 370

412 Jorg P. Vogele

TABLE 3 Infant mortality in the large towns of Prussia (more than 200,000 inhabitants), 1875-1910 (per100 births)

BerlinBreslauCharlottenburgCologneDortmundDusseldorfDuisburgEssenFrankfurt a MHannoverKielKomgsbergMagdeburgNeukollnStettin

1875-80

27 727 929 123618 219 918 818 416 316 415 724 923 8

-25 6

Legitimate

1881-90

24 727 426 924 816 919 919 317 716 217 116 725 524 5

_27 0

1891-00

20 524 518.823.217 719 719 717 013 817 016 623 623 825 729 7

1901-10

16 721 012 818 815 816 016 414 412 313 614 717 120 417 623 8

1875-80

47 745 155 928 235 436 129 541 531 928 828 452 037 9

-42 5

Illegitimate

1881-90

42 543 448 232 038 339 931 342 831 027 029 848 845 1

-445

1891-00

38 138 039 936 838 539 845 141 031 430 232 143 243 453 047 6

1901-10

26 330 326 231 943 331 737 537 228 224 727 533 532 337 437 8

Source Taschenbuch des Statistischen Amis der Stadt Dusseldorf 5 vermehrte Auflage (Dusseldorf,1913), p 23

registered a substantially lower proportion of illegitimate births when com-pared to Prussia. In the Regterungsbezirk only 3.41 per cent of all live birthswere illegitimate, in the Rhemprouinz 3.78, and in the whole of Prussia 8.38.48

In the city of Dusseldorf, however, the rate was considerably higher than inthe surrounding areas, in 1909 there were 8.8 per cent illegitimate infants inthe city, 4.0 in the Regiemngsbezirk, 4.2 in the Rheinprovinz, 7.8 in Prussia and9.0 in the German Empire.49

However, it remains difficult to estimate the extent to which this risk grouphad reduced chances of survival. In areas with a high proportion of illegitimateinfants they were most likely to be socially accepted to a larger degree andcorrespondingly found better living conditions than in areas with a low pro-portion. Despite the substantially higher mortality rates of illegitimate infants,the impact of the legitimacy composition of births on the level of the overallinfant mortality rate remained rather limited. For example, in 1905 there were8,912 infants born in Dusseldorf, among them 8,679 live births.50 Of thelatter, 7,995 (92.1 per cent) were legitimate and 684 (7.9 per cent) illegitimate.The total number of infant deaths during this year amounted to 1,658, ofwhich 1,414 were legitimate and 244 illegitimate births Legitimate and il-legitimate infant mortality was 17.7 and 35 7 respectively; on the whole there

48 Cewerbestatistik von Preussen, 3 Teil D e r R e g i e r u n g s b e z i r k Dusseldorf , 2 B a n d ( Iser lohn,1865), p 97

49 Source Taschenbuch des Statistischen Amts der Stadt Dusseldorf 3 v e r m e h r t e Auflage (Dusse l -dorf, 1911), p 17

50 Source Jahresbencht des Statistischen Amts der Stadt Dusseldorffiir 1905, pp 3 - 5


were 19.1 infant deaths per 100 births. Assuming identical death rates, but atwice as high proportion of illegitimate births, the overall infant mortalityrates would have risen merely by about 1.4 per cent, from 19.1 to 20.5.51

Infant mortality and feeding practices

Feeding practices had a more obvious impact on the survival chances ofinfants. Different mortality levels from digestive diseases were determined toa large extent by regional variations in infant feeding methods. Artificialfeeding was associated with high infant mortality, and extensive breastfeedingwith low infant death rates. Breastfeeding minimizes the chances both ofmalnutrition and the acquisition of infectious diseases. Breast milk is not onlyconsidered to be nutritionally ideal, it is also clean and transmits immunitiesfrom the mother to the child. On the other hand, artificial food promotesbacterial infection, especially when prepared with milk, water, or, as wascustomary in some areas, with food pre-chewed by an adult. Processing andstoring usually did not meet the requirements of hygiene. Furthermore, duringthe nineteenth century, artificial food often could not sufficiently supply pro-teins and vitamins. A substantial survival advantage of breastfed infants overartificially nourished infants, particularly in the first months of life,52 has beennoted for historical populations as well as for contemporary less developedcountries.53 In Sweden at the beginning of the nineteenth century, for example,low infant mortality rates were linked with legislative coercion to breastfedinfants.54 High infant mortality rates in the eastern and south-eastern parts ofGermany, on the other hand, were associated with the fact that breastfeedingwas relatively uncommon in these regions. The potential impact of breastfeed-ing on the survival chances of infants becomes obvious, if a homogenous socialgroup is analysed, in this case 628 women, who gave birth in the Colognematernity asylum (Wochnerinnen-Asyl). Only poor married women wereadmitted, and the municipal welfare committee had to confirm their indi-gence. Among this group, in 1900-01 infant mortality fluctuated between 8deaths per 100, when breastfeeding was practised for over three months, and37 for infants who were not breastfed at all or only for a period shorter thanthree months.55

The wide spatial variation in mortality from diseases of the digestive sys-tem, however, cannot be attributed solely to differences in infant feedingpractices, but must also reflect the role of local environmental factors. Feedingpractices, for example, cannot explain the extensive mortality differences inthe western towns of Germany, where deaths from intestinal diseases per

51 See also Kintner, 'Determinants', pp 55-652 J. Knodel and H Kintner , ' T h e Impact of Breastfeeding Pat terns on the B i o m e t n c Analysis

of Infant Mor ta l i ty ' , Demography, 14, 4 (1977), 3 9 1 - 4 0 953 P n n z i n g , Handbuch (1906), pp 2 9 0 - 4 , Knodel and Kintner , ' Impact o f Breastfeeding'54 Selter, 'D ie N o t h w e n d i g k e i t der Mut te rb rus t fur die E r n a h r u n g der Saughnge ' , Centralblalt

fir allgememe Cesundheitspftege, 21 (1902), 3 7 7 - 9 2 , p . 38955 Dietr ich, 'Saughngse rnahrung und W o c h n e n n n e n - A s y l e ' , Centralblatt fir allgememe Gesund-

heuspfiege, 21 (1902), 4 6 - 5 3 , p 48

414 Jbrg P. Vbgele

10,000 ranged from 29.15 in Frankfurt, over 58.15 in Dusseldorf, to 78.88 inCologne during 1877, although infant feeding patterns were essentially simi-lar. Infant mortality in Cologne was actually higher than in the non-breast-feeding area of Breslau, although both towns had similar birth rates. Hereenvironmental factors might have played a more important role. Furthermore,the decline in urban infant mortality went along with a decline in breastfeedingpractices in the large towns of Germany.56 For example in Berlin, in 1885 55.5per cent of all infants were exclusively breastfed, 4.0 per cent additionally fedwith animal milk, and 33.9 per cent solely raised on animal milk. In 1900 only31.4 per cent were breastfed, 0.7 per cent supplied with additional animalmilk, whereas 54.8 per cent relied on animal milk.57 In the traditional non-breastfeeding areas the percentage of infants nourished solely artificially waseven higher. According to material from the Munich children's hospital, thenumber of infants never breastfed was 78.3 per cent in 1861-69, rising evenhigher in the following decades to 82 3 (1870-78) and 86.4 (1879-86).58 Inview of decreasing breastfeeding in the last quarter of the nineteenth century,environmental improvement must have played a more important role indetermining the decline in infant mortality. Improved conditions of hygienenot only restricted outbreaks of severe typhoid fever that still occurred in the1880s and early 1890s, but also the nutritional adequacy of artificial food wasimproved, thereby reducing the survival advantage of breastfed infants overthose artificially nourished.59

Infant mortality and sanitary reform

Sanitary reform focused on infrastructural measures that were intended toimprove the hygienic conditions of the urban environment. Particular em-phasis was placed on improved water supply and sewerage systems, but alsoselectively included disinfection, control of food and especially municipal milksupply 60 Illness and disease were no longer regarded as natural and thereforea matter of destiny, but rather as something that could be positively influencedby improving environmental conditions. Specific measures were carried outin an increasingly systematic manner during the last decades of the nineteenthand the early twentieth century. At the turn of the century public health

56 Selter, ' N o t h w e n d i g k e i t ' , pp 3 8 2 - 5 , M Hohlfeld, ' U e b e r den U m f a n g der nat i i rhchenSaughngsernahrung in Leipzig ' , Deutsche Medizimsche Wochenschnft, 31 (1905), 1391-4 , p 1392,Pnnz ing , Handbuch (1906), p 294, H J Kintner , 'Trends and Regional Differences in Breastfeed-ing in G e r m a n y from 1871 to 1937', Journal of Family History (1985), 163-82, pp 169-72

H N e u m a n n , ' U b e r die Haufigkeit des Stillens', Deutsche Mediamsche Wochenschnft, 28(1902), p 795

58 F Biiller, 'Ursachen und Folgen des Nichtst i l lens in der Bevolkerung M u n c h e n s ' , Jahrbuchfur Kmderheilkunde, 16 (1887), 3 1 3 - 4 0 , p 320, T Eschench , 'D ie Ursachen und Folgen desNichtsti l lens bei der Bevo lke rung M u n c h e n s ' , Munchener Mediamsche Wochenschnft, 34 (1887),233-5 , 256—9, p 233, H Seidlmayer, Geburtenzahl, Saughngssterblichkeit und Sttllung in Munchen wden letzten 50Jahren (Munich, 1937), p 29

59 J D W r a y , ' M a t e r n a l Nu t r i t ion . Breast-Feeding and Infant S u r v i v a l , in W H Mosley . ed ,Human Nutrition and Reproduction ( N e w York , 1978), 197-230, pp 2 1 3 - 1 9

60 T Weyl , 'Assamerung ' , in T Weyl , ed . Soziale Hygiene Handbuch der Hygiene, 4Supplement -Band (Jena. 1904), 1-27, p 1


measures had become a fixed component of municipal budgets. Munich, forexample, spent 16.3 per cent of its 1894 budget on public health issues.61

A major factor affecting the urban environment during the second half ofthe nineteenth century was the installation and expansion of central watersupply and sewerage systems. Following the pioneering example of Britain,sanitary reforms spread in Germany during the last decades of the nineteenthcentury, with major building activity for central water supply systems beingundertaken in the large towns particularly in the 1870s and 1880s; by 1900 allthe large towns had a central water supply. The expansion of sewerage sys-tems followed the construction of water-works at a delayed interval; the mainperiod of construction occurred around the turn of the century.

The level and trend of the death rate from typhoid fever is often consideredto be a classic test for assessing the health-impact of these reforms.62 In viewof the declining frequency of breastfeeding in urban areas, the supply of cleanwater was especially important. Artificial food was to a large extent preparedwith water, and animal milk was also often diluted. This implies that thepotential effect of these infrastructural public health measures should also bemeasurable in relation to the subsequent development of death rates fromdigestive diseases as well as infant mortality.

The limitations and successes of infrastructural measures can be demon-strated using the example of Hamburg. The city had an early central watersupply (1842), yet failed to install a filter plant. This deficiency cruciallycontributed to the devastating effect of the infamous cholera epidemic of 1892which caused 8,616 deaths Hamburg was the only large German city to beaffected by this epidemic.63 Figure 3 reveals that infant mortality reached amaximum during this year. After the installation of a filter plant in thefollowing year not only did the risk of dying from cholera and typhoid feverdecrease,64 but also deaths from digestive diseases were reduced, increasingthe survival chances of infants (Figures 3 and 4).65 Recent research indicatesthat sanitary reform had an impact on urban infant mortality.66 In general,however, the quantitative and qualitative state of central water supply andurban sewerage systems left much to be desired.67 In the nineteenth century

61 K Singer, Die Abmmderung der Sterblichkeitsziffer Munchens Em Beitrag zur Frage hygiemscherund sozialpolitischer Maassnahmen aufdie Cesundheit der Stadte (Munich , 1985), p 19

62 R O t t o , R Spree and J Vogele 'Seuchen und Seuchenbekampfung in deutschen Stadtenwahrend des 19 und friihen 20 Jah rhunde r t s . Stand und Desiderate der Forschung ' , Medtzmhis-tonsches Journal, 25 (1990), 286-304 , pp 297-301

63 SeeR Evans,DeathinHamburg Society andPoliticsintheCholeraYearsl830-l910(Ox(or&,\9%l)64 J Vogele , ' T h e U r b a n Mor ta l i ty Decline in G e r m a n y , 1870-1913 S o m e Prel iminary Re -

sults ' , in G Kearns, W R Lee, M C Nelson and J Rogers , eds , Improving the Public HealthEssays in Medical History (Liverpool , for thcoming)

65 Source Verojfentlichungen des Kaiserlichen Cesundheitsamtes (Beilagen), (1878—1914)66 J C . Brown, 'Public Health Reform and the Decline in Urban Mortal i ty T h e Case of

Germany, 1876-1912' , in G Kearns, W R Lee, M C Nelson and J Rogers, eds , Improvingthe Public Health Essays in Medical History (Liverpool, forthcoming), Stockel, 'Saughngssterbhch-keit', 2 3 0 - 1 , J Vogele, 'Sanitare Reformen und der Wandel der Sterbhchkeitsverhaltmsse indeutschen Stadten, 1870-1913' , VSWC, 80 (1993) H 3, 345-65

67 Vogele, 'Urban Mortal i ty '

416 Jbrg P. Vogele

Death rate (per 10,000 living)

600

500

300

200

• I 1 1 I

A1

<\

i i < i

Af i

' \

• i i

a

' *^/V

'i'i'i' i' i' i1 i' i' i

V °

a

V

I860 1885 1890 1895

Year

1905 1915

""""Hamburg "•'Towns > 15.000 inh

FIG 4. Mortality from acute digestive diseases in German towns 1877-1913

they were still generally restricted to the large towns. Numerous outbreaks oftyphoid fever in these places reflected the persistent risk of using drinkingwater. Correspondingly, the average death rate from acute digestive diseasesremained at a permanent high level in towns over 15,000 inhabitants duringthe last decades of the nineteenth century (Figure 4).

Sanitary reforms also included the establishment of municipal milk depots.For Britain it has been argued that the substantial fall in infant mortality afterthe turn of the century could be attributed mainly to the improved provisionof pasteurized milk, the introduction and popularization of dried milk as an in-fant food, as well as the widespread use of condensed or evaporated milk 68

Recent studies are more sceptical about the impact of municipal milk supplyon the decline of infant mortality in Britain As breastfeeding remained wide-spread, the health of weaned infants was more dependent on relative poverty,the mother's education and her participation in the labour force, on over-crowded housing, environmental conditions, and a general decline in fer-tility.69 Others suggest that consumption of often qualitatively poor milkcontributed to ill-health,70 particularly because of its impact on infant deathsfrom tuberculosis and digestive diseases.

68 M W Beaver, 'Population, Infant Mortality and Milk', Population Studies, 27 (1973), 243-54, D Dwork, War is Good for Babies and Other Young Children A History of the Infant and ChildWelfare Movement w England, 1898-1918 (London/New York, 1987)

Woods, Watterson, and Woodward, 'Causes', pp 116-2070 P J Atkins, 'White Poison' The Social Consequences of Milk Consumption in London,

1850-1939', Sofia/ History of Medicine, 5 (1992), 207-28, p 227


In view of the decline in breastfeeding in German towns and in areas wherebreastfeeding was already only practised to a minor extent, sanitary reformersregarded the quality of milk as a matter of supreme importance in the battleto combat the highly feared summer diarrhoea. However, whereas the townsspent substantial sums of money on central water supply and sewerage sys-tems, backed by official arguments concerning diseases such as cholera andtyphoid, the supply and distribution of municipal milk only began with somedelay and hesitation. As contemporary doctors complained, high infant mor-tality rates were regarded as somewhat inevitable,71 even though some of themain reasons for these high rates were already known. Contemporaries indi-cated that differences in infant mortality between urban and rural areas couldbe attributed to higher mortality in towns during the hot summer months.72

This was substantially supported by the prevalence of digestive diseases in thecause-of-death panorama. Indeed, the hot period of the year claimed numerousvictims among infants. In Berlin in 1885, for example, more than 45 per centof all infant deaths occurred in June, July, and August.73 Increasingly fewerexperts attributed this to urban housing conditions.74 Instead, the case foradequate infant nutrition became more and more the focus of attention. Theterm children's milk (Kindermilch) was created,75 and the supply of animal milkwas considered to be a municipal task of the highest priority.

In Germany the sale of pasteurized or sterilized milk commenced in the late1880s with the invention of the so-called Soxhlet apparatus (1886), by whichmilk was heated within the bottle. Complete sterilization, however, wasexpensive, and therefore rarely applied in practice. Heating the milk to orbeyond its boiling point, of course, caused the loss of a considerable amountof vitamins. Pasteurization of milk, discovered in the 1860s by Louis Pasteur, re-mained controversial, since it was taken as the cause for various infant diseases,for example infant scurvy (Sauglmgsskorbut). It was therefore suggested thatpasteurized milk should only be used in the hot summer months, with the specialtreatment of milk regarded as superfluous in the cooler periods of the year.76 Milkwas not among the subjects dealt with in the nutrition laws (Nahrungsmittelgesetz),and in 1900 only three towns (Berlin, Dresden, and Munich) had special regula-tions concerning children's milk. In general, the quality of milk remaineddubious. Partly sterilized milk was sold in green or brown bottles in order to makevisual quality control impossible for the consumer.77

71 Pfaffenholz, 'Wichtige Aufgaben der offenthchen und pnvaten Wohlfahrtspflege auf demGebiet der kunsthchen Ernahrung der Saughnge', Centralblatt fir allgemeitie Gesundheitspflege, 21(1902), 393-416 , pp 4 0 2 - 3

72 Schlossmann, 'S tudien ' , p 13773 W u r z b u r g , 'Saughngss terbhchkei t ' , p 7474 For an overview see C Flugge, Groflstadtwohnungen und Kleinhaussiedluiigen m ihrer Emwir-

kung auf die Volksgesutidhat (Jena, 1916)Pfaffenholz, 'Saughngs-Sterblichkeic und Kindermilch', Centralblatt fir allgemeitie Gesundheits-

pflete, 21 (1902), 183-200, p 183L Spiegel, 'Kommunale Milchversorgung', Schriften des Verems fir Socialpolitik, 128 (1908),

219-43, p 23177 Flugge, 'Aufgaben', p 321

418 Jorg P. Vogele

Municipal milk was increasingly based on an imported supply, as cattle-breeding in the towns declined rapidly. In the early years the supply andpreparation of milk was completely in the hands of private enterprises and freetradesmen. By the late-nineteenth century, by contrast, many municipalitieshad established their own management system dealing with milk supply anddistribution.78 In the first decade of the twentieth century, however, only ahandful of towns had established municipal milk depots. Yet, many othertowns also recognized the importance of regulating milk supply and insistedon certain standards. As a consequence quality improved In Diisseldorf, forexample, in 1895, 60 out of the 265 official examinations of milk were unsatis-factory, in 1906 only 125 out of 3,743 failed to pass the test.79 Police ordi-nances (Pohzeiverordnungen) by 1901 included regulations about the quality andthe importation of milk from outside the town. Merchants who wanted to sellmilk had to register at the police station. Each consignment of milk had to besealed before being imported, and its origin clearly identified, so that in thecase of irregularities the responsible merchants could be easily found.80

Bacteriological examination, however, remained difficult and was hardlyfeasible in practice,81 given the expense involved. It was for this reason thatthe amount of specially treated and controlled milk for children available inthe towns remained very limited, often amounting to only 500 litres in a townof 100,000 inhabitants.82 Sold at a price of 50-60 Pfennige per litre this typeof milk was only an option for well-established circles of the population.Elements of the food industry clearly were, at least at times, not in favour ofmunicipal milk supply and promoted their own artificial products. There werereports from Bonn in 1902 that families, immediately following a child's birth,received a brochure signed by a paediatrician, promoting the use of powderedinfant food, the so-called children's flour (Kindermehl).83

Contemporaries were convinced of the success of municipal milk supply interms of improving the state of health of infants and young children. How-ever, demographic data confirms the generally poor state of municipal milkprovision. An adequate and widespread milk supply should have contributedto a decline in mortality from tuberculosis and digestive diseases amonginfants, as well as weakening the climatic structure of infant mortality byreducing the extent of its seasonal fluctuation In the German towns, however,the death rates from tuberculosis among infants actually increased from 29(per 10,000 births) in 1877 to 44 in 1900 and then declined only slightly to 38by 1907 (Table 2), although it fell in the age-groups from 5 years onwards.Death rates from digestive diseases, on the other hand, showed a strongdecline at the beginning of the twentieth century when assessed from average

78 Spiegel, 'Milchversorgung', p 23279 F Schrakamp, 'Gesundheitswesen', in T Weyl, ed , Die Assamenmg der Stadte in Emzeldar-

stellungen, Vol 2 2 Die Assamerung von Dusseldorf (Le ipz ig , 1908), 8 3 - 1 1 9 , p 11080 Schrakamp, 'Gesundheitswesen', p 110-1381 Pfaffenholz, 'Aufgaben', p 40082 Pfaffenholz, 'Aufgaben', p 404, Spiegel, 'Milchversorgung', p 22983 Cramer, Contribution to the discussion following Pfaffenholz, 'Aufgaben', p 419


figures for all towns with a population exceeding 15,000 inhabitants (Figure4) and in the ten largest towns (Table 2). However, climatic dependencyremained; the cold summer of 1902 was accompanied by low infant mortality,84

but the hot summer of 1911 registered high infant mortality rates.85 Seasonalfluctuations persisted, with a pronounced summer peak (Figure 5),86 andaffected most powerfully artificially fed infants.87 Excellent long sources inHamburg show that summer infant mortality was even higher in the twentiethcentury, when compared with the preceding hundred years (Figure 6).88 InBerlin, the summer peak reached its pinnacle in 1885, yet in 1900 it was stillmuch higher than in the period 1850-74, and did not disappear completelyuntil 1926-28.89 A similar development can be observed in Munich, wherethe summer peak did not vanish until the second decade of the twentiethcentury,90 in part as a result of increased breastfeeding,91 Whereas urban infantdeath rates were already below the corresponding rural figures after 1900,higher urban infant mortality in the hot summer months remained a commonfeature. Figure 7 shows the actual infant mortality rates in urban and ruralareas of Prussia at the beginning of the twentieth century.92 Demographicevidence, therefore, confirms the view that a qualitatively satisfactory milksupply reaching a substantial portion of the population, and especially thepoor, had still not been developed by the first decade of the twentieth century.Contemporary attempts to provide adequate infant food, however, demon-strate increasing concern about high infant mortality rates. Declining birthrates raised fears about the nation's future in economic and military respects,and led to a concentration of forces in the battle against high infant mortality.

84 See Figure 1, Spiegel, 'Mi lchverso rgung ' , p 225; W Kruse, 'Was lehren uns die letztenJahrzehnte und der heisse Somrae r 1911 uber die Saughngss terbhchkei t und lhre Bekampfung ' ,Centralblatt fur allgememe Cesundheitspflege, 31 (1912), 175-201 , p 179

85 Kruse, 'Was lehren uns ' , pp 175-201 , Seidlmayer, Geburteiizahl, p 20 A m o r e elaboratestatistical test, howeve r , is difficult to achieve due to inadequate source material S o m e informa-tion about climatic condi t ions in the t o w n s of the sample is only available for a selection of thesetowns for a small n u m b e r of years from the Statisttsches Jahrbuch Deutscher Stadle, 1 (1890) o n w a r d sCorre la t ing average annual air tempera ture with annual overall morta l i ty , infant mor ta l i ty anddeath rates from various diseases in the years 1888-96 leaves us wi th a rather weak associationT h e relatively s t rongest , and surpris ingly negat ive correlation is achieved wi th mor ta l i ty fromdigestive diseases (r = - 0 5200, sig = 0 000, n = 67), followed by infant mor ta l i ty (r = - 0 4375,sig = 0 000, n = 67) A correlation using tempera ture m a x i m a leaves an even weaker relation Forthe period 1907-11 correlations are weaker Here the s t rongest interrelat ionship is be tween annualaverage t empera tu re and deaths from tuberculosis o f the lungs (r = - 0 5212, sig = 0 000, n = 49)

86 Source H Silberglen, 'Kindersterbl ichkei t in europaischen Grosss tadten ' , Huitieme CongresInternational D'Hygiene El De De'mographte, T o m e VII (Budapest , 1896), 4 4 3 - 5 6 , p 454

87 A E Imhof, 'Untersch iedhche Saughngss terbhchkei t ' , p 352^Sources Statistisches Handbuch fur den Hamburgischen Staat (1920), p 48, p 70; Die Cesundheitsver-

hallmsse Hamburgs im neunzehntenjahrhundert Den arztlichen Thetlnehmern der li Versammhmg Deut-scher Naturforscher und Arzlegewtdmel von dem Medicinal-Collegium ( H a m b u r g , 1901), p 108, p 114

89 Imhof, 'Saughngssterbl ichkei t ' , p 35290 R Hecker , 'Studien uber Sterbhchkei t , Todesursachen und Ernah rung M u n c h e n e r Saug-

hnge ' , Archw fur Hygiene, 93 (1923), 280 -94 , p 2871 Seidlmayer, Ceburtenzahl, p 22, p 31

92 Source W Kruse , 'Was lehren uns die letzten Jahrzehnte und der heisse S o m m e r 1911 uberdie Saughngss terbhchkei t und lhre Bekampfung ' , Centralblatt fur allgememe Cesundheitspflege, 31(1912), 175-201 , p 178

420 Jorg P. Vdgele

Per cent

-°- Dusseldorf "T Cologne ^ Frankfurt/M * Berlin ^Breslau "0" Dresden A Leipzig & Munich

FIG. 5 Seasonal distribution of infant mortahty (%) 1889/93

Per cent

Quarter

-°-1850 +1860 ^1870 * 1880 X i 8 9 0 O 1900 A 1910 X"1911

FIG. 6 Seasonal distribution of infant mortality in Hamburg (%)

Infant mortality and living standards

Many contemporaries in fact attributed high infant mortality rates primarilyto the participation of women in the labour force. High female labour forceparticipation was associated with an absence of breastfeeding and a general



35

+Urban "• Rural

FIG. 7 Seasonal distribution of infant mortality in Prussia, 1911 (per 100 births)

neglect of child care.93 The effects of female labour force participation on thehealth of infants, however, remain ambiguous and differ through time andspace. Some contemporaries were already puzzled by the negative relationbetween female earnings and the infant mortality rate.94 For the sample ofGerman towns there was no statistical correlation between the percentage ofwomen not participating in the official labour market in 1907 and infantmortality rates (1901-13): r = -0.1106, sig (two-tailed) = 0.761.95 Pre-sumably, improved health conditions increasingly reduced the differences inmortality between breastfed and artificially nourished infants. Against thisbackground the working mother's wage, which bought a better standard ofliving for the family, might have become a more important factor in thesurvival chances of infants.96 After the turn of the century generally higherfemale wages for day labour reduced infant mortality in towns: there is arelatively strong statistical correlation between female wages and infant mor-tality (r = —0.6974, sig (two-tailed) = 0.025),97 indicating again that forfamilies at the lowest end of the income distribution additional income out-weighed the impact of reduced breastfeeding. The lack of correlation between

93 C Dyhouse , 'Work ing-Class M o t h e r s and Infant Morta l i ty in England, 1895-1914 ' , Journalof Social History, 12 (1978), 248 -67 , p p . 2 5 1 - 3

94 W o o d s , Watterson and W o o d w a r d , 'Causes ' , p 11595 Sources Statistik des Deutschen Reichs, N F 207 (1907), Veroffentlichungen des Kaiserhchen

Gesundheitsamtes (Beilagen), (1901-1914)96 Dyhouse , 'Working-Class Mothers ' , pp 254-797 Sources Stalistisches Jahrbuch Deutscher Stadte, 19 (1913), p 826, Verojfemlichungen des Kaiser-

hchen Gesundheitsamtes (Beiiagen), (1901-1914) See also Brown, 'Public Health'

422 Jorg P. Vogele

female labour force participation and infant mortality, however, could well bethe result of regional differences, which possibly disguised potential inter-relations. Whereas in the highly industrialized western areas of Germany,differences in infant mortality rates between breastfed and artificially-fed in-fants increasingly vanished, they remained evident in the less industrializedeastern areas. Industrialization, therefore, reinforced contemporary regionaldifferences.98 Towns of the sample belonging to traditionally non-breastfeeding areas, i.e. Breslau, Munich, and Nuremberg, registered a highproportion of female labour force participation as well as high infant mortalityrates. In these towns high female labour force participation could not counter-balance low living standards, so that the absence of breastfeeding had a strongernegative impact on the survival chances of infants than in the better-off moreindustrialized western areas In Eastern Prussia, for example, the growth ofJunker estates based on grain monoculture led to a peasant class which neededto exploit the labour of its female members in order to subsist This resultednot only in high adult female mortality, but also in high infant mortality rates,caused by the inability of many mothers to breastfeed. In Western Prussia, onthe other hand, land reforms in the early nineteenth century resulting in awider distribution of peasant land ownership, greater crop variations and morelivestock, brought about better health conditions."

In southern Germany the impoverishment of the vast majority of the popu-lation as a consequence of the Thirty Years' War continued into the followingperiods. Regions which suffered most deeply from the Napoleonic Wars 1796-1809 registered high infant mortality rates. Battles, troop movements, har-vest failure, the requisition of cattle, as well as war contributions led to severepauperization. 10° In addition, those regions based on agriculture were hitseverely by the spread of potato-rot. Yet, even in periods of bad harvest andhunger crisis, corn was exported to neighbouring foreign countries.101 Finally,a peasantry poor in purchasing power made the survival of small-scale tradersin the towns difficult. All this and in particular the lack of cattle led tonotorious infant feeding practices using a thick and sweetened meal-papDummies were often filled with mashed rusk and sugar Even poppy-seedextracts and opium were added as a sedative. As a result of this poor situation,Imhof claims that the local population developed a fatalistic mentality, parti-cularly in Catholic areas, based on a deeply rooted 'system of wastage ofhuman life'.102 This, however, remains dubious in view of the fact thateconomic growth and industrialization from the late 1860s, the expansion of

98 Lee, 'Germany', p 15699 W R Lee, 'Mortality Levels and Agrarian Reform in Early 19th Century Prussia Some

Regional Evidence', in T Bengtsson, G Fndhzius and R Ohlsson, eds , Pre-Indiistnal PopulationChange The Mortality Decline and Short-Term Population Movements (Stockholm. 1984), 161-90.pp 168-72

100 Pnnzing, 'Entwicklung'. p 598101 J Vogele, Getreidemarkte am Bodensee mi 19 Jahrhundert Strnkturen und Enlwicklungen (St

Kathannen, 1989), p 37102 Imhof, 'Saughngssterbhchkeit'


the railway network, and numerous good harvests, led to a decline in infantmortality. Indeed, this decrease started even earlier than in the northern partsof Germany.103

Germany's accelerated industrialization from the late 1860s and early 1870s'onwards brought about a rise in living standards. Between 1871 and 1913 realwages almost doubled and the average growth of the economy was signifi-cantly higher than in England.104 This rise was distributed almost equally overthis period, and may have had a substantial impact on the decline of mortalityvia improved nutrition. Although it is difficult to obtain direct evidence ofthis for the sample of towns, 1OD the congruence between rising real wages andfalling death rates in towns and cities is impressive. As a supporting effect thegrowing world market in corn, which reduced prices,106 provided betternutritional provision, independent of rising income.107 Recent research hasindeed indicated an improvement in diet during the nineteenth century, fromwhich the lower classes increasingly benefited in the second half of thecentury 108 From the 1870/80s onwards the estimated actual calorific intakesurpassed the recommended amount.109 Despite this, infant mortality rates inPrussia did not decline among working class families before the end of thenineteenth century, with the consequence that class-specific differences ininfant mortality levels increased.110 Only the twentieth century has broughtabout a decline in infant mortality among all social groups in Prussia. Thisclearly indicates that additional factors, such as cultural or mental aspects,might have played a strong part in determining levels and trends of urbaninfant mortality, ft also indicates that further research ought to refine theindicators under discussion As this, however, is hardly feasible in a macro-

103 Pnnzing, 'Entwicklung', p 599KM -p j Orsagh, 'Lohne in Deutschland 1871-1913 Neue Literatur und weitere Ergebmsse',

Zeitschnfi fur diegesamte Staatswissenschaft, 125 (1969), 476-83, p 483 Germany —1 3 per cent p aEngland —0 9 per cent

More direct evidence could be obtained by analysing infant mortality according to parentaloccupation for the sample of towns, following R Spree's analysis of Prussian figures See Spree,Health and Social Class, pp 63-71, Spree, 'On Infant Mortality', See also Stockel, 'Saughngssterb-lichkeit', pp 240-9 Incorporating adequate material on occupational mortality for the sample ofthe ten largest towns would require an analysis of local data sources, which are scatteredFurthermore, the number of records to be collated and analysed would by far surpass the scopeof the present macro-level approach

106 Vogele , Getretdemarkte, p 39107 R A Dickler , 'Labour Marke t Pressure Aspects o f Agricul tural G r o w t h in the Eastern

Region of Prussia, 1840-1914 A Case Study of Economic-Demographic Interrelations during theDemographic Transition', unpublished Ph D thesis, University of Pennsylvania (1975), pp 177-8

108 W Abdel , Stufen der Ernahrung Erne histonsche Skizze (Got t ingen , 1981), H J Teu teberg ,'Studien zur Volksernahrung unter sozial- und wirtschaftsgeschichthchen Aspekten', in H JTeuteberg and G. Wiegelmann, eds , Der Wandel der Nahniiigsgewohnheiten unter dent Emflufi derIndustnalisierung (Gottingen, 1972), 21-221, H J Teuteberg, 'Der Verzehr von Nahrungsmittelnin Deutschland pro Kopf und Jahr seit Beginn der Industnalisierung (1850-1975) Versuch emerquantitativen Langzeitanalyse', Archw fiir Soztalgeschuhte, 19 (1979), 331-88, H J Teuteberg andG Wiege lmann , Unsere taghche Kost Ceschichte und regtonale Pragung (Muns te r , 1986)

109 W G Hoffmann, Das Wachslum der deutschen Wirtschaft sen der Mute des 19 Jahrhunderts(Berlin, 1965), p 659

110 Spree, Health and Social Class, pp 63-71, Spree, 'On Infant Mortality'

424 Jorg P. Vogele

level approach, further analysis should be directed to the use of disaggregatedmicro-level data.

Conclusion

The analysis indicates that in the long run the urban-industrial world has hada positive impact on health conditions. By creating health-preserving environ-mental conditions and through a rising standard of living, urban populationslost their earlier disadvantage in survival chances when compared to ruralareas. This went so far that even traditional risk factors were largely elimi-nated. Despite a decrease in breastfeeding in towns, infant mortality ratesstarted to decline earlier and more vehemently in urban areas in comparisonto the national average. By implementing an increasing number of measuresin the following decades, a health-preserving system was created, whichcounterbalanced traditional urban health risks, as well as those created byindustrialization. These measures included, for example, the expansion of thehealth sector in general, municipal housing, statutory sickness insurance aswell as increasing access to infant care (zugehende Saughngsfursorge). The impactof the last two factors, in particular, seems to have been underestimated. Withrespect to sickness insurance the focus has been on the medical treatment ofthe sick worker, neglecting a major objective of the scheme which was toprovide financial support in case of sickness. Therefore, in the case of sicknessof the main wage earner of the family, it was possible to maintain a standardof living for the whole family which prevented it from falling immediatelyinto poverty. Thus, sickness insurance had a broader impact and could havesubstantially contributed to the decline in infant mortality.

Health-securing measures, created in the period under investigation, formedthe essential base for a health-preserving way of life. The role of the health-care recipients, however, was of equal importance. The provision of satisfac-tory hygienic milk, for instance, was of little value, if the purchasers did notkeep it cool, if they stored it in the warmest place in the house, ' ' ' and neitherwashed their hands before preparing the milk nor cleaned the bottles after use.In other words, the potential benefits of an improved health-related infrastruc-ture remained underutilized in the absence of a corresponding improvementin personal hygiene. The evolution of a collectively rooted personal hygieneregime was therefore a prerequisite. This in turn was the result of an increasein rationalist conduct of life, which was based on the assumption that onecould mould one's own destiny. Corresponding mentalities and norms ofbehaviour pervaded society from top to bottom. Within this process medicineincreasingly took over educating and controlling tasks. As already mentioned,nineteenth-century doctors, for the most part, remained excluded from accessto the sick infant. Medical guides and handbooks of infant care required a

111 F Soxhlet, 'Ueber Kindermilch und Saughngs-Ernahrung', Munchener Meduimsche Wochen-schnft, 33 (1886), 253-6, 276-8, p 255


certain level of education and therefore did not receive a wide circulation anddistribution. Nineteenth-century breastfeeding campaigns remained withoutbroad effect.112 With the beginning of the twentieth century, however, doc-tors realized the necessity of an active health-care approach to the population.From 1907 onwards they developed the so-called advancing infant care whichenabled medical personnel to be in direct contact with mother and childthrough house-visits.113 This made the doctors transmission vehicles for thesenew values, which successfully spread among working-class people andaffected contemporary views on hygiene and health.

112 Stockel, 'Saughngssterbhchkeit', 252-6113 Engel and Behrendt, 'Saughngsfursorge', pp 28-194, F Rott, 'Die drei Senkungspenoden

der Saughngssterbhchkeit lm Deutschen Reiche', Gesundheitsfirsorge fir das Kmdesalter, 2 (1926/27), 491-508, F Rott, 'Gesundheitsfursorge fur das Kmdesalter', in M v Pfaundler and ASchlossmann, eds , Handbuch der Kttiderheilkunde Em Buchfar den praktischen Arzt, Vol. 1 (Berlin,1931), 4th edn , 187-252

Urban Infant Mortality in Imperial Germany

Documents