In ancient times physicians wrote primarily on the care of infants, and only incidentally about chil­dren’s diseases, because their concept of medicine stressed the maintenance of health rather than the diagnosis of specific disease entities (for medical perspectives on children during antiquity, see Etienne 1973).

The earliest of these “pediatric” texts known to us was that of Soranus of Ephesus (active around A.D. 100), On Gynecology, which in­cluded 23 chapters on infant care (see Soranus 1956; also Ruhrah 1925; Still 1931; Garrison 1965; Peiper 1966).

Soranus was a leader of the Methodist sect at a time when Greek medicine was enlivened by various contending schools of thought. Methodism taught that disease was due to excessive relaxation or con­traction of internal pores of the body, leading to immoderate secretion and moisture in the first in­stance and to diminished secretion and dryness in the second. The cause of disease was considered un­important, stress being laid instead on treatment that, crudely put, consisted of inducing the contrary state, drying the moist or humidifying the dry. In his section on infant management, Soranus concen­trated on the practicalities of care and treatment without slavish adherence to the tenets of Method­ism. The result was a pragmatic guide uncompli­cated by theoretical or speculative overtones.

During the second century, Claudius Galen inau­gurated a radical change in perspective by setting out his own complex theoretical synthesis. In so do­ing he established humoral theory, already several hundred years old, as the main guide to understand­ing health and disease. He was so successful that for the next 1,500 years humoral doctrine permeated most medical writings, including those on children and their diseases (for a guide to humoral theory, see Ackerknecht 1968).

According to this doctrine, the four elements, earth, fire, air, and water, were related to four quali­ties, hot, dry, cold, and wet.

Book 1 of Galen’s Hygiene, entitled “The Art of Preserving Health,” contains five chapters on the care of the newborn and young child. Milk, accord­ing to Galen, was the ideal food for infants “since they have a moister constitution than those of other ages” (Galen 1951). Infants were also warm by na­ture, and “in the best constituted bodies,” this qual­ity would remain constant until adolescence, when it would increase in intensity. In contrast, the normal body would steadily dry out from infancy onward to reach desiccation in old age. In addition, there were individual differences, inborn divergences from the ideal constitution, that would require special man­agement, but these were not specifically discussed in Galen’s Hygiene.

However, the concept of constitutional change dur­ing growth and development, associated with that of innate differences in temperament or constitution, gave great flexibility to humoral theory and an al­most limitless explanatory power. In part because the theory was so intellectually satisfying, its valid­ity was not seriously questioned until about the six­teenth century. Even after the surfacing of doubts, humoralism survived, in various eroded forms, into the nineteenth century. In contrast to the modern ontological concept of diseases as specific entities (each with its own cause, natural history, and cure), under humoral doctrine illness was considered a per­sonal manifestation of humoral imbalance due to environmental stress interacting with the individ­ual’s own special constitution.

The interplay between the humoral and ontologi­cal interpretation of disease can be illustrated by an examination of medieval and early modem descrip­tions of measles and smallpox. Rhazes, an Arab phi­losopher and physician at the turn of the tenth cen­tury, wrote a text on children, Practica puerorum, unusual for the times in that it dealt only with diseases and ignored infant management (for an En­glish translation and analysis, see Radbill 1971). Rhazes also penned a treatise on smallpox and mea­sles, which he distinguished between but considered to have the same cause: putrefaction and fermenta­tion of the blood. Whether a person exhibited the symptoms of smallpox or those of measles depended on humoral constitutions (see Rhazes 1939).

Until about the seventeenth century, physicians followed Rhazes' example in describing the two dis­eases under the same heading. Thomas Phaire, a Welsh physician whose pediatric text, published in 1545, was the first to appear in the English lan­guage, began his chapter on smallpox and measles thus: “This disease is common and familiar... it hath obtained a distinction into two kinds: that is to say, varioli the measles, and morbilli, called of us the smallpox. They be both of one nature, and proceed of one cause, saving that the measles are engendered of the inflammation of blood, and the smallpox of the inflammation of blood mingled with choler [yellow bile]” (Phaire 1965).

In the following century an English physician, Thomas Sydenham, provided the first description of measles per se, as well as of scarlet fever and “chorea” (St.

Nosology did not prosper in pediatrics, where eighteenth-century physicians continued to list symptoms and diseases unsystematically. Yet it was one of Sydenham’s pupils, Walter Harris, who first diverged from classical humoral theroy. In his Trea­tise of the Acute Diseases of Infants (published in 1698 in Latin) Harris insisted that the prevailing cause of illness in infancy was excess acidity. Cure involved neutralizing and absorbing the acid with “testaceous powders” such as powdered oyster shell, crab’s claws, egg shell, chalk, coral, mother of pearl, and burned ivory. The residue was then removed by purgation. Harris recommended rhubarb, a mild laxative by the standards of the time.

For the next 100 years, until the early nineteenth century, Harris’s acid theory was predominant. George Still (1931) has suggested that its popularity was due to its simplicity. The old humoral pathology no longer seemed a reliable guide: it had become very complex because commentators felt free to amend it to suit themselves. In contrast, Harris sup­plied his readers with a simple cause and remedy for most early childhood ailments.

The late eighteenth century witnessed a revolu­tion in chemistry.

In the second half of the eighteenth century some medical men, notably Giovanni Morgagni and John Hunter, stressed the need to correlate clinical signs during life with later autopsy findings. This ap­proach was occasionally used in the eighteenth­century pediatrics, by Robert Whytt, for example, in his observations on dropsy in the brain or acute hydrocephalus (reprinted and discussed in Ruhrah 1925). Whytt gave an excellent clinical account of this condition, which, however, would not be recog­nized as meningitis due to tuberculosis until the early nineteenth century. The regular practice of performing autopsies began with the advent of chil­dren’s hospitals, which were introduced in Paris af­ter the French Revolution. The Enfants Malades was established in 1802 for sick children over the age of 2 years and the ancient foundling hospital, the Enfants Trouves, was reorganized to accommo­date ailing babies as well as unwanted ones.

From contemporary reports one gathers that physi­cians were rarely refused permission to perform au­topsies (Crosse 1815). With a relatively large num­ber of children under their care, hospital doctors could also apply the so-called numerical method, a rudimentary statistical analysis by which mortality as well as complication and recovery rates could be assessed (Ackerknecht 1967). Under this system cer­tain diseases, such as early childhood tuberculosis, could now be recognized and reported. Whytfs “dropsy of the brain” was classified as a manifesta­tion of tuberculosis when it was discovered that the disease was associated with granular tubercular de­posits on the membranes covering the brain (Papa- voine 1830).

French pediatricians attempted to classify chil­dren’s diseases, whereas, as already mentioned, ear­lier authors had dedicated chapters with complete impartiality to symptoms (e.g., nightmares, sneez­ing, hiccoughs) and to specific morbid conditions (e.g., mumps, measles, smallpox). In 1828 Charles M. Billard published a text exclusively on infant diseases, which he classified according to the site of the main lesion: skin infections, those of cellular tissue, of the digestive tract, of the respiratory and cardiac systems, of nervous or cerebrospinal origin, of the locomotor and generative systems, and finally congenital disorders. Freddric Rilliet and Antoine Barthez, authors of a general pediatric treatise first published in 1843, used a different method. Diseases were classified primarily according to underlying pathology: phlegmasias (inflammations), hydropsies (accumulation of watery fluid), hemorrhages, gan­grenes, neuroses, acute specific diseases, the various kinds of tuberculosis, and finally entozoan (para­sitic, caused by various worms) diseases. Under these headings, further subdivisions were made ac­cording to the site affected (Barthez and Rilliet 1853).

In early-nineteenth-century Paris, with children’s hospitals growing in number and size, pediatrics emerged as a specialty inasmuch as physicians there could engage full time in the care of children. They published extensively, stimulating interest in pediat­ric research abroad, although in the United States and in Britain physicians continued throughout the nineteenth century to be generalists rather than specialists (for the history of pediatrics in the United States, see Cone 1979). At first, the concept of pediat­ric hospitals did not gain approval in English­speaking countries for a variety of reasons, some of which are remarkably consonant with modern opin­ion. It was said that children, especially babies, should not be separated from their parents and would have a better chance of recovering at home than in hospitals, where, as evident from the Pari­sian reports, cross-infection was rife (Anon. 1843). On Malthusian and moral grounds, Henry Broug­ham held that “the gratuitous maintenance of poor children, may safely be pronounced dangerous to society, in proportion as it directly relieves the par­ent from his burthen. It removes the only check upon improvident marriages, and one of the principal guards of chastity” (Brougham 1823). With Broug­ham, foundling hospitals were the prime but not the exclusive subject of contention. Nevertheless, by the 1850s opinion had changed sufficiently to allow for the establishment of pediatric hospitals, albeit small ones initially, in the United States and Britain. One reason for the change was pressure from physicians who argued that the intensive investigation of pedi­atric disease required a hospital base (Anon. 1849).

For the next hundred years, pediatric hospitals and clinics served as sites of further investigation and treatment of disease. After World War II, however, the earlier concept that babies and small children should not be separated from their mothers surfaced again with renewed intensity. Psychoanalysts in par­ticular drew attention to the long-term emotional effects of maternal deprivation, most evident in chil­dren reared in orphanages but also demonstrable in babies hospitalized for acute disease or for surgery. Most influential perhaps was the 1951 report pre­pared by John Bowlby (1951) for the United Nations program on the welfare of homeless children, but as early as 1942 Harry Bakwin, a pediatrician, had drawn attention to apathy in babies confined to hospi­tals. Indeed, from the mid-nineteenth century on­ward some physicians had opposed the isolation of infants, but the dictates of hospital organization and fears of cross-infection severely restricted parental access to their children until the psychoanalytic con­cept of maternal deprivation was elaborated in the 1950s.

Cross-infection had frequently ravaged pediatric wards in the early nineteenth century; hence, with the discovery that microorganisms caused conta­gious diseases, the isolation of potentially infective patients seemed an obvious solution. As Harry Dowl­ing (1977) points out, between 1880 and 1900, 21 microorganisms were identified as the causes of spe­cific diseases, including the bacilii responsible for tuberculosis, typhoid fever, and diphtheria. Also dis­covered were the pneumococcus and meningococcus associated with the common forms of pneumonia and meningitis, and the streptococcus whose hemolytic form produced scarlet fever and rheumatic fever. Yet knowing the causes of these illnesses was not imme­diately helpful. More obvious benefits would have to await specific vaccines and antibiotics.

Babies continued to die at a steady rate. As histori­ans have frequently indicated, in England the infant mortality rate did not decline during the nineteenth century, hovering between about 140 and 160 per 1,000 from 1839, when national vital statistics were first recorded, until 1900, even though all other age groups exhibited a falling death rate from midcen­tury onward (Logan 1950; McKeown 1976; F. B. Smith 1979; Winter 1982; Wohl 1983; Dwork 1987).

Most other countries that kept vital statistics were looking at similar trends. In Massachusetts, the earliest state to keep continuous records, the infant mortality rate was similar to that of England. During the second half of the nineteenth century, Massachusetts reported the highest infant mortality rate of 170 for the years 1870 to 1874 and the lowest of 123 for the years 1855 to 1859; for 1895 to 1899 the rate was 153 (U.S. Bureau of the Census 1960). The French and Germans reported even higher in­fant death rates; in 1895 the French infant mortality rate was 177, the German one 230 (Mitchell 1975). However, the Scandinavian countries and Ireland were doing much better, with lower and falling in­fant death rates during the second half of the cen­tury, demonstrating that much loss of life elsewhere was unnecessary. In 1895 the infant mortality rate in Norway was 96, in Sweden 95, in Denmark 137, and in Ireland 104 (Mitchell 1975). In England, re­gional statistics showed that infant mortality was often twice as high in industrial areas as in rural ones, which, together with the Scandinavian experi­ence, suggested that predominantly agricultural so­cieties exhibited conditions most favorable to infant survival (for contemporary discussions, see Jones 1894; Newman 1907). Industrial areas were more lethal, it was often concluded, because so many moth­ers went out to work, abandoning their babies to casual care and bottle feeding. Dirt and environmen­tal pollution, including food contamination, were seen as secondary to the damage caused primarily by a lack of breast feeding.

According to English vital statistics, the main kill­ers ofinfants were “atrophy” and debility, pulmonary diseases (bronchitis and pneumonia), convulsions and meningitis, diarrheal diseases, and tuberculosis (Newman 1907). In the United States the picture was similar, except that diarrheal diseases accounted for a larger porportion of deaths, while pulmonary dis­eases were less prominent than in England (Cone 1976). By the last third of the nineteenth century, smallpox, which formerly had been particularly haz­ardous to babies, no longer accounted for many deaths, probably owing to widespread vaccination in the first months of life. But otherwise, medical prog­ress, including the recognition of microorganisms as the cause of infectious diseases, seems to have had little relevance to the plight of infants. The diarrheal diseases, for example, included a variety of conditions that defied simple classification. Job Lewis Smith, a noted New York pediatrician, discussed simple diar­rhea, intestinal catarrh or enterocolitis, the dreaded cholera infantum, enteritis, and colitis in the 1890 edition of his Treatise on the Diseases of Infancy and Childhood. Infantile cholera, the most acute and se­vere type, was so called because its symptoms were so similar to those of epidemic cholera. Diarrheal dis­eases were more common and fatal in the hot summer months, but the hot weather alone could not be re­sponsible because summer diarrhea caused less harm in rural areas. Smith therefore subscribed to the time-honored view that the cause was related “to the state of the atmosphere engendered by heat where unsanitary conditions exist, as in large cities.” Once, out of deference to germ theorists, he sent intestines from a child who had died of cholera infantum for examination by William H. Welch. The report was inconclusive because all kinds Ofbacteria were found on the surface of the intestines. Still, Victor C. Vaughan, professor of hygiene and physiological chemistry at the University of Michigan, was confi­dent that the diarrheas were caused by toxin­producing bacteria. In his opinion, “there is not a specific micro-organism, as there is in tuberculosis, but any one or more of a large class of germs, the individual members of which differ from one another sufficiently morphologically to be regarded as dis­tinct species, may be present and may produce the symptoms” (Vaughan 1897).

Recent historians, particularly Anthony S. Wohl (1983), have pointed out that maternal malnutrition probably contributed significantly to the high infant mortality. Chronically undernourished women gave birth to puny, sickly babies ill-equipped to withstand exposure to infection and the other hazards prevail­ing in a working-class environment, such as cold and damp housing and an inadequate diet. On the whole, however, Victorian physicians did not consider pov­erty and its consequences to be responsible for the unacceptably high infant mortality rate. Instead, they focused on the dangers of artificial feeding; influenced by middle-class expectations and stan­dards, they berated women for going out to work, leaving their infants to be bottle-fed and possibly drugged by ignorant baby minders. It apparently did not occur to them that a mother’s earnings might be essential for family survival, probably because a working mother was so contrary to the general Victo­rian image of womanhood.

By the onset of the First World War, infant mortal­ity had fallen significantly in most European coun­tries and in the United States. The infant mortality rate for Massachusetts dropped from 141 in 1900 to 116 in 1914 (Woodbury 1926). In England and Wales the reduction was from 154 in 1900 to 105 in 1914. Some of this abrupt change must be attributed to factors that had emerged in the nineteenth century but had taken time to prove beneficial, for example, better nutrition for the population at large leading to the birth Ofhealthier babies (for discussions of the reasons for the decline of mortality at the turn of the century, see Woodbury 1926; McKeown 1976; Dy- house 1978; F. B. Smith 1979; Dwork 1987). The contribution of direct medical intervention to the initial decline in mortality remains debatable, but progress in the understanding of disease began to exert observable effects on infant survival after World War I.

So numerous were the twentieth-century break­throughs in understanding pediatric disease that only a brief summary can be attempted here. A list­ing of important disease categories is accompanied by a short discussion of how changes in traditional thinking gradually provided a new basis for reme­dial action.