Nutritional aspects of human lactation*

HUMAN LACTATION Nutritional aspects of human lactation* A. M. THOMSON 1 & A. E. BLACK 2 This paper reviews the literature on the incidence and duration of breast-feeding in various countries, the volume and composition of breast milk, the health and nutrition of breast-fed babies as judged by growth and morbidity, maternal nutritional requirements during lactation, and the effect ofprolonged lactation on maternal health. It appears that lactation can be as well sustained by impoverished as by affluent mothers, and that even in communities where malnutrition is common the average growth of infants is satisfactory up to the age of about 3 months on a diet of breast milk alone. Breast milk appears to have specific anti-infective properties, but prolonged breast-feeding will not prevent infections among older infants reared in a poor environment. The authors believe that breast-feeding is the best form of nutrition for the young infant and deplore its decline in modern industrial societies. The recommendations of various FAOQWHO Expert Groups on nutritional intakes during lactation are summarized. The needfor an increased daily energy intake of 4.2 MJ (I 000 kcal) is questioned, and an increase of2.5 MJ (600 kcal) is suggested. Data on the effect ofprolonged lactation on the health of the mother are scanty; body weight appears to be maintained even among poorly nourished mothers. The authors stress the need for well-planned and technically adequate studies of the material and psychological factors involved in breast feeding. IS BREAST-FEEDING BEST? It is difficult-perhaps impossible-to discuss breast-feeding without appearing to take sides. That being so, it is necessary to state our own belief that, in general (but not necessarily always), breast-feed- ing is best. If it is not, the mammary organs of the human female must be regarded as interesting relics of an inefficient evolutionary process for feeding the young. We deplore the fact that, even in the " per- missive society ", most western mothers regard breast-feeding in public as immodest and in private as inconvenient, but we recognize that the organiza- tion of industrial society and of life in nuclear families creates practical difficulties. The fact that modern technology has provided substitutes for breast milk that many mothers and doctors find acceptable and satisfactory does not impair our belief in the basic biological " rightness " of breast- feeding sustained by adequate lactation. * From the Medical Research Council Reproduction and Growth Unit, Princess Mary Maternity Hospital, Newcastle upon Tyne, NE2 3BD, England. ' Director. 2 Technical Officer. That belief finds support in an enormous amount of medical literature. For example, a recent sym- posium on " The uniqueness of human milk " (45) advocated breast-feeding on the grounds of the considerable evidence in its favour. Such advocacy has a long history. Toverud et al. (91) quote Oliver Wendell Holmes (1809-1894): " a pair of substan- tial mammary glands have the advantage over the two hemispheres of the most learned professor's brain in the art of compounding a nutritive fluid for infants." Such views, however, do not necessarily reflect the less vocal opinions of many medical practitioners, and they have been counteracted, very successfully, by the appeal of advertising coupled with the pressures of life in modern industrial society and the introduction of new values into traditional societies. For the purpose of this review, our concern with breast-feeding and lactation is scientific. We accept Popper's (70) definition of science, according to which the statement " Breast feeding is best" is a hypothesis that has to be tested by the most stringent means available. We recognize that breast-feeding has already lost most of its popularity in western society and is losing ground in many developing 3336 - 163 - BULL. WORLD HEALTH ORGAN., Vol. 52, 1965 A. M. THOMSON & A. E. BLACK countries. We do not know how this tendency is to be halted, let alone reversed. But assuredly the cause of breast-feeding will not be assisted by failure to examine scientific evidence as critically as possible, or by basing advocacy on insufficient, misleading, or erroneous information. INCIDENCE AND DURATION OF BREAST-FEEDING Breast-feeding and lactation are not simply differ- ent words for the same thing. Breast-feeding is a form of human behaviour for which lactation is necessary, but the amount of milk produced need not correspond to either the productive capacity of the mother or the nutritional needs of the baby. It should therefore be remembered that data about breast-feeding do not necessarily reflect the ade- quacy of lactation. The Annex summarizes, without pretensions to completeness, some recent reports on the incidence and duration of breast-feeding in various countries. Some of the samples may be highly selective, and comparisons are complicated by the many different ways in which the data are presented. Some general features are nevertheless apparent. As Solien de Gonzales (82) and others have pointed out, breast-feeding tends to be the most common and sustained the longest in " traditional " communities and in those with the fewest economic advantages. It is the least common and the least well sustained in relatively affluent urbanized communi- ties, where most babies may be artificially fed from birth. This broad inverse correlation between the incidence of breast-feeding and economic and educa- tional status ceases to be valid in some communi- ties-e.g., the more affluent and educated classes of North America-where breast-feeding may be more common than in poorer classes (11, 34, 74). Then, however, there seems to be a tendency for supple- mentary foods to be introduced at an early stage and for weaning from the breast to take place after a few months at most. Where data over a period of time exist they show that, wherever mothers have been exposed to the culture of modern industrial society, the incidence of breast-feeding has been declining. For example, in the United Kingdom, Hytten (38), showed that the proportion of infants still being breast-fed at 3 months of age fell from 60-80% in 1929 to less than 30% in 1955. Similar trends have been reported from Poland (13), from USA (62), and from Guate- mala (82). There may have been some reversal of the downward trend among more educated women .(see below), but this is unlikely to have had any marked effect on the general tendency. Since breast-feeding is a form of human be- haviour, it is important to study the motivation behind it. There seems to be little evidence that women have a strong " instinct " to breast-feed, despite the physical and emotional satisfaction that breast-feeding is often claimed to give. Prothero (71) considered that, in a group of London mothers, the establishment and maintenance of breast-feeding was a matter of intellectual conviction rather than of instinctive urge. In Aberdeen, Scotland, very few of a group of primiparae found breast-feeding to be completely simple, satisfying, and satisfactory (39). The steady decline of breast-feeding in most com- munities, despite much medical propaganda in its favour (not, however, outweighing commercial and other propaganda on behalf of breast milk substi- tutes), is a fact that requires further study. It does not suffice merely to hope that modern technology and communications will fail to penetrate " primi- tive" populations where sustained breast-feeding continues to be the usual form of maternal be- haviour. There appears to be no scientific information on possible ethnic (genetic) differences in the capacity to lactate and breast-feed. Our own experience suggests that these may nevertheless be important: in West Africa, breast-feeding seems to be not only almost universal but also highly adequate and easy (60), whereas in the United Kingdom early weaning is common and many women seem unable to secrete enough milk, or milk of sufficient energy content, to satisfy the needs of the baby (37). The shape of the "equipment ", i.e., breasts and nipples, may be different in different races. VOLUME AND COMPOSMON OF BREAST MILK Reliable information on the volume of milk pro- duced is scanty. It is not easy to measure daily milk production by manual or mechanical expression; most estimates have been based on test-weighing the baby itself-a procedure that necessitates interfer- ence with " natural " breast-feeding, especially where feeding is undertaken casually or on demand. An- other difficulty is that the amount of milk secreted depends to a great extent on the demand-i.e., on the frequency and vigour of suckling. Reports of late failure of lactation are thus particularly difficult to interpret, possibly because in many instances the 164 HUMAN LACTATION child may have been put to the breast much less frequently than during early lactation. Sen6cal (79) reported the following 24-h milk outputs, measured by test-weighing in hospital, in a West African community where suckling often con- tinues with little abatement for 18 months or more: for babies aged less than 3 months, 4-6 months, 7-9 months, 10-12 months, 13-18 months, and over 18 months, the average yields were 583, 659, 648, 685, 738, and 702 ml, respectively. Sustained yields, averaging more than 500 ml during the first year, have also been reported for poor Indian women by Rao et al. (73) and by Rajalakshmi (72). There is little comparable information from western com- munities; the extremely high yields reported for late as well as for early lactation in Detroit (66, 67) are by no means representative, since the subjects were professional wet-nurses. Disregarding these results, it appears from information summarized by Mor- rison (64) that established milk yields of 600-700 ml per day may be typical-i.e., not much bigger than those reported from developing countries. Assuming an average energy value for mature breast milk of 0.3 MJ (70 kcal) per 100 ml, a daily yield of 700 ml of milk represents an energy supply to the baby of nearly 2.1 MJ (500 kcal). The data of Fomon et al. (29) suggest that babies aged about 3 months need about 0.4 MJ (95 kcal) per kg of body weight per day, and the average weight of babies of that age is about 5.5 kg. On this basis, it appears that a diet of breast milk alone should be sufficient for babies up to about 3 months of age. After that, supplementation is likely to become increasingly necessary if the energy needs of the baby are to be fully met. The composition of human milk has been de- scribed and compared with that of cow's milk by Macy & Kelly (56), Gyorgy (35), and, with special reference to tropical communities, by Bailey (4). The effects of maternal nutrition on the composition of milk were discussed by Hytten & Thomson (41). Here, only a few points of special interest need to be summarized. (a) The amount of protein in human milk per 100 ml or protein energy per 100 kcal (0.42 MJ) appears to be lower in human beings than in faster- growing mammals (8). The concentrations above 12 g/litre that are sometimes reported in the litera- ture have usually been obtained by applying a conversion factor to total nitrogen, including non- protein nitrogen. Human milk has a relatively low protein level, but it has a higher level of S-containing amino acids (with a high cystine : methionine ratio) and of many essential amino acids and histidine (52) than cow's milk has. (b) Human milk contains more long-chain fatty acids and free fatty acids than cows' milk does. Its fatty acid composition tends to reflect that of fats in the maternal diet and may be affected by the state of the maternal energy balance (42, 53). (c) There appears to be general agreement that the output of breast milk and the concentration in breast milk of protein, fat, carbohydrate, calcium, and iron are little influenced by the nature and amount of the maternal diet within a remarkably wide range of intakes, and that the nutritive value of the milk is well sustained in prolonged lactation. In New Guinea, Bailey (4) found the proximate composition of milk to be much the same after 18-24 months of lactation as it was at 6-12 months. Karmarkar et al. (48) assert that supplementation with protein and fat of the diets of very poor Indian women in Baroda had no effect on the yield but some beneficial effects on the protein or fat content of breast milk; on the other hand, Gopalan (30) in South India found that extra dietary protein caused an increase in milk yield associated with some reduction of protein concentra- tion. Even under famine conditions, the effect on breast milk appears to be slighter than might have been expected, probably because maternal tissues can be used to subsidize lactation. The effect of the maternal diet on human milk yields seems to be much slighter than is the case with farm animals; but it should be remembered that such animals have usually been bred to produce much greater yields of milk than are required by the offspring. (d) The amount of fat-soluble vitamins contained in breast milk is influenced to some extent, and that of water-soluble vitamins to a great extent, by the nature of the maternal diet. We shall not attempt to deal with the complex reasons for the cessation of lactation, but it may be noted that pregnancy has been described as a " com- mon cause for cessation of breast feeding " in a poor Indian community (86). In a study of 120 poor-class Egyptian women, Kamal et al. (47) also reported pregnancy during lactation to be the commonest cause of weaning. They say that " It is evident that pregnancy did not cause milk insufficiency, for it was not a major cause of supplementation... Thus it appears that endogenous steroids do not have a pronounced effect on lactation, so long as suckling is 165 A. M. THOMSON & A. E. BLACK maintained ". The first part of that quotation, at least, would seem to require independent confirma- tion, since it seems to contradict the widespread belief that pregnancy " competes " with lactation. However, pregnancy does not necessarily lead to weaning. Cantrelle & Leridon (12) say that in rural Senegal 30% of mothers conceived while breast- feeding; 10% were still nursing when 4 months pregnant, and 1.5% during the last month of preg- nancy. The literature thus gives the impression that the babies of mothers who keep on breast-feeding will outgrow the supply of milk, but that, with continued suckling, the volume and nutritive value of the milk remains at a high level for a long time, even during a subsequent pregnancy. That view, which is based on rather flimsy evidence, needs to be reassessed. It is possible that mothers who stop breast-feeding at a relatively early stage are those whose milk supply is failing. Data from Nigeria (W. Z. Billewicz & D. Morley, unpublished observations, 1970) show that, irrespective of the age at which it took place, the weaning of Nigerian children in Ilesha was accom- panied by a considerable but temporary decrease in the rate at which weight was being gained. This phenomenon did not occur among rural Gambian children (59). It seems difficult to explain the differ- ence in terms of an effect that occurs, or fails to occur, after weaning. A more probable hypothesis is that, in Nigeria, the loss of weight was due to failing lactation, which preceded weaning. In the Gambia, the effects of failing lactation would usually be mitigated by the local custom of passing the baby to -other lactating women within the family, if the mother was unable to satisfy it. There is a considerable need for objective studies of the natural history of breast-feeding and asso- ciated phenomena in developing as well as in " ad- vanced" communities. Existing information is scanty, usually crude, and often biased by the atti- tude of the observer. THE HEALTH AND NUTRITION OF BREAST-FED INFANTS As has been shown above, theoretical calculations suggest that a diet of breast milk alone will become insufficient for the average growing child when it attains a weight of about 5.5 kg, around the age of 3 months. In practice, supplementation of some kind has already been introduced by many mothers. The nutritional problem thereafter centres progressively on the amount and nature of the supplementary food, and the importance of breast milk gradually diminishes. The interpretation of data on the growth and health of infants from the point of view of sustained breast-feeding is complicated by other factors. The introduction of supplementary foods, which may be of poor nutritive value, is accompanied by increasing exposure to bacteria and parasites at the very time when the immunity acquired passively before birth is dwindling. The resulting illness may affect appetite and the ability to suckle, thus having a potentially adverse effect on the mother's ability to lactate as well as on the growth and wellbeing of the child. Conversely, a poorly-nourished child may have an increased susceptibility to infections and infesta- tions. Infections Breast milk is commonly considered to have spe- cific anti-infective properties by comparison with breast milk substitutes. According to a review by Mata & Wyatt (58): (a) there is "an incomplete understanding of the significance of antibodies in human colostrum and milk, although no doubt exists of their participation in host defense, parti- cularly derived from their presence in the gastro- intestinal lumen "; nevertheless it is reasonable to consider that such antibodies " act importantly in control of invading agents "; (b) a human milk diet leads to a bifidobacteria flora that is probably antagonistic to certain pathogens; and (c) breast milk seems to have anti-infective properties through other factors, such as a "resistance factor " to staphylococcus aureus, lysozyme, complement, inter- feron, and cells with immunological capacity. Mata & Wyatt conclude that " Human milk is definitely more effective than cows' milk in protecting the newborn infant against infection ". There is no reason to doubt that conclusion, in so far as it applies to the neonate andyoung infant, but whether breast milk has an important anti-infective role in older infants appears to be a more open question. In an important Swedish study, Mellander et al. (61) note that "Up to the 1930's most investigations showed a considerably higher in- cidence of infant infections in infants reared by the bottle. Critical analysis of these studies, however, clearly reveals that poor standards of hygiene and concomitant increased risks of infection were of great significance... and it is therefore difficult to determine what effect the differences in composition 166 HUMAN LACTATION of the artificial mixtures and breast milk may have had. Recent investigations, in which more attention has been paid to these sources of error, have still in most cases shown a preponderance of acute infec- tions in bottle-fed infants... but the differences have often been only moderate ". In their own investigations, during which great care was taken to ensure comparability between infants fed in different ways, a small but statistically significant excess of upper-respiratory-tract infections and acute diar- rhoea was found in infants that had been breast-fed for less than 2 weeks compared with those given 61/2 months or more of " plain breast feeding ". This suggests that, where the hygiene of the environment is satisfactory, the influence of breast- feeding on the incidence. of infection in infants is marginal, and this is consistent with the important fact that infant mortality rates have fallen to remark- ably low levels in most " advanced " countries de- spite the rapid decline of breast-feeding. Kubat et al. (54) and Bouchalova' & Omelka (9) in Czechoslova- kia, and the Royal College of General Practitioners (75) in the United Kingdom failed to find any advantage of breast-feeding over bottle-feeding in infants and young children. Few, if any, similarly controlled studies have been undertaken in countries with unfavourable environ- mental conditions. In such circumstances, fully breast-fed infants have a low incidence of diarrhoeal and other infections and infestations during the early months of life, but meet with a tide of infection during the latter part of infancy (60, 77). It also appears that earlier weaning is being accompanied by a reduction in the ages at which serious forms of malnutrition are found (26). In Malaysia, Dugdale (18) found the type of feeding to have no statistically significant effect on the frequency of minor respiratory or alimentary illnesses in infants. In Senegal, Cantrelle & Leridon (12) found, contrary to their expectation, that child- ren who died between the ages of 2 and 4 years had been breast-fed for longer (25.6 months on average) than children who survived (24.3 months); " a pos- sible explanation may be that breast-feeding is often prolonged for a child whose health is delicate, although this does not always prevent its death some time after it has been weaned ". Certainly there is no reason to believe that main- tenance of prolonged breast-feeding, without atten- tion to the hygiene of the environment and to the improvement of infant care generally, will be effec- tive in reducing the prevalence of infections, or to doubt that, where the environment is favourable, bottle feeding can be accompanied by a satisfactorily low incidence of infection. Improved methods of supplementary feeding, applied in isolation, appear to be relatively ineffective. Edozien (19) working in Nigeria with children aged 12-36 months, found increased rates of growth when milk protein sup- plements were given, but mortality remained high unless malaria and other infections were dealt with. Becroft & Bailey (6) in New Guinea and Venkata- chalam et al. (93) in India failed to show that nutritional supplementation of breast-fed infants im- proved growth. Similarly, in Guatemala, Behar et al. (7) failed to demonstrate a striking effect of supplementary feeding on disease incidence or on growth and development. The same workers (78) pointed out that " Precise knowledge of breast milk intake is woefully lacking for representative developing regions. The value and necessity of breast-feeding in most developing regions for the first six months of life is firmly established, in some places to such an extent that it is almost a matter of ' nurse or die '. Is excessively long breast-feeding, however, actually harmful because of the assumption that the child is being well fed in later months when in reality he is not? Or is it a saving grace because of social and economic inability to provide a better substitute? These are questions largely unanswered. " Other forms of morbidity Breast-feeding may be of importance as regards factors other than infection. Replacing it by substi- tutes with a high content of saturated fatty acids, such as full-cream cows' milk, may lead to defective absorption of calcium and to greater risks of hypo- calcaemic tetany in infants (96). Grulee & Sanford (33) reported that breast-fed infants had a relatively low incidence of infantile eczema; this does not appear to have been confirmed. Nor does there appear to be confirmation of a report by Osborn (69) that teenagers who had been entirely breast-fed for 2 months or more had less coronary disease than those who had been bottle-fed since birth. Davies (16) has quoted Osborn in support of his theory that cows" milk protein is an antigen that can lead to atheroma and coronary thrombosis. Paediatricians in some countries are concerned about hyper- natraemia caused by feeding over-strength cows' milk formulas, and the possible early onset of obesity due to concomitant excessive calorie intake (10). The problem of lactose intolerance among bottle-fed 167 A. M. THOMSON & A. E. BLACK Table 1. Growth increments of fully or predominantly breast-fed infants from birth to 3 months of age Head Mean Wegtgi .eghgi circum-Place and source birtheht Weightgain Lengthgain ference Sample and type of feeding(kg) (mm) gain (mm) Growth standards (65) M 3.40 2.32 98 56 Boston children of north European stock F 3.36 2.26 93 53 Iowa, USA (28) M 3.44 2.57 99 - 104 Caucasian infants > 2.5 kg birth weight. Breast-fed F 3.24 2.25 97 - plus supplements UK (63) M 3.40 2.56 - - 60 infants 3.2-3.6 kg birth weight. Fully breast-fed at 15, F f16, or 17 weeks Keneba, Gambia M 2.9 2.9 116 - 99 infants in rural village, poor economic status and very(I.A. McGregor et al., un- F 2.7 2.4 108 - high prevalence of infectious diseases published observations, 1968) Dakar, Senegal (20, 57) M 3.29 3.23 108 564 356 infants > 2.5 kg birth weight, mostly breast-fed F 3.14 2.71 101 514 Jamaica (84) M 3.48 2.43 - - 229 children, total births in rural survey area; poor eco- F 3.31 2.19 - - nomic status. Almost all breast-fed during first 3 months Hyderabad, India (86) M 2.8 2.25 1015 575 87 infants. Rural sample, poor economic status. Breast- F 2.7 1.80 885 490 feeding usual for over 1 year Bombay, India (46) M 2.84 1.72 1024 538 61 infants of the sweeper community, very poor; 90% F 2.73 1.78 820 569 fully or mainly breast-fed at 3 months children is not within the scope of this review, but is may be noted that lactose intolerance has been suggested as a cause of persistent diarrhoea and wasting among breast-fed children in Africa (50). Growth Growth rates provide a useful means of summing up the general nutrition and health of breast-fed infants. It was mentioned above that the volume and proximate composition of breast milk appear to be little affected by the nutritional status of the mother. This is strongly confirmed by records of the growth of infants whose diet consisted exclusively or pre- dominantly of breast milk. It is, of course, necessary to consider growth rates during the first three or four months of life only, since after that the breast milk supply cannot be expected to meet the complete nutritional needs of the infant, and more or less adequate supplementary feeding is usually provided. Table 1 gives differences between size at birth and size at 3 months, as shown in the well-known Harvard growth standards (65), together with incre- mental data from a number of fully longitudinal or semilongitudinal growth studies of infants stated to be fully breast-fed or coming from communities where breast-feeding is the rule. It seems reasonable to believe that the increments shown in the table were sustained for all practical purposes by a diet of breast milk alone. The figures show that early growth in weight, length, and head circumference is seldom seriously defective, and often well " up to standard " even when the mothers are from the poorest classes in Bombay, impoverished peasants living near Hyderabad, or hard-working women in one of the unhealthiest localities in West Africa. The evidence in Table 1 could, of course, be greatly augmented by using published data from cross-sectional growth studies from most countries of the world. There is practically universal agreement that breast milk is satisfactory as the sole food of young infants, ir- respective of the standards of nutrition and health that prevail in the communities where the mothers live. (This is not the same thing as saying that individual mothers who are ill or badly nourished or overworked always continue to breast-feed satis- factorily; they often do not. We are concerned here with average patterns in communities, and they are unquestionably satisfactory.) That satisfactory situation does not continue. In communities with slender resources, poor hygiene, low standards of nutrition, unsatisfactory levels of education, and so on, it is usual for the growth of infants to falter severely towards the end of the first 6 months. This cannot be interpreted simply in terms 168 HUMAN LACTATION Table 2. Growths increments in breast-fed babies of differing birth weights a Increments Measurement at 3 months as Birth weight (birth-3 months) % of birth measurement group head head(kg) weight length circum- weight length circum-(kg) (mm) ference (%) (%) ference (mm) (%) < 2.0 2.44 926 559 233 120 118 2.0-2.5 1.97 866 455 183 118 114 2.5-3.0 2.09 886 458 174 118 113 3.0-3.5 2.07 895 472 163 118 114 3.5 + 1.79 819 433 147 116 112 a Data of Banik et al. (5). of inadequate lactation or failure to breast-feed; as already indicated, the situation is complicated by the necessity to provide supplementary foods and by exposure to infection at the time when passively acquired immunity is fading. Whether it is true, as many have said, that prolonged breast-feeding is particularly important where mothers are unable or unwilling to provide satisfactory breast milk substi- tutes under hygienic conditions during the critically important transition to an ordinary mixed diet, still seems to require objective confirmation. Low birth weight In view of the special vulnerability of infants of low birth weight (95), it is important to consider the extent to which breast-feeding is adequate for such infants. There is ample evidence that, in general, breast-fed infants of low birth weight gain less during early life than those that are artificially fed (1), which suggests that breast-feeding may not be best in those special circumstances. The situation is complicated by the fact that many underweight babies are fed for a considerable time on expressed or banked breast milk by means of a tube or bottle; whether this is really equivalent to breast-feeding may be worth investigation. Banik et al. (5) describe growth during the first 3 months of life of Indian (Delhi) babies most of whom were, presumably, exclusively breast-fed. Table 2 is a summary of their data. Whether the increments in the weight of low-birth- weight babies are to be regarded as satisfactory or not, these data suggest that lactation is not seriously impaired, if at all, in mothers who have given birth to babies of low birth weight. MATERNAL NUTRITIONAL REQUIREMENTS DURING LACTATION Table 3 summarizes the recommendations of various Expert Groups convened by FAO and WHO. To facilitate comparisons, the data have been reduced to a common basis, which is specified in footnotes. These recommended allowances include a margin of safety (except for energy), which may be large: they are not minimum physiological require- ments. Nevertheless, they show that nutritional re- quirements during lactation are usually higher than those during pregnancy, which in turn are higher than those of a comparable nonpregnant woman. The FAO Second Committee on Calorie Require- ments (21) considered that lactation imposed an additional requirement of 4.2 MJ (1000 kcal) per day on the following basis. The average milk production was taken as 850 ml per day, with an energy value of 2.5 MJ (600 kcal); this energy was tentatively as- sumed to be provided from food energy with a conversion efficiency of 60%. The FAO energy allowance does not correspond with the extra intake of healthy women who are lactating successfully, and it is certainly too high. Thomson et al. (90) found that the average dietary energy intake of 23 Scottish lactating women was 2.47 MJ (591 kcal) per day greater than that of 32 women who were bottle- feeding their babies. Taking into account probable subsidization of lactation through loss of maternal body fat, they calculated that the average efficiency of human milk production was greater than 90%, in agreement with findings, nearly 40 years earlier, of Shukers et al. (81). On this basis, according to Thomson et al., " for a woman producing 500-600 169 A. M. THOMSON & A. E. BLACK Table 3. Daily allowances recommended by FAQ/WHO expert groups for nonpregnant, pregnant, and lactating women a Nonpregnant Pregnant Lactating energy, J (kcal) b 9.6 MJ (2 300 kcal) 10.5 MJ (2 500 kcal) 13.8 MJ (3 300 kcal) protein, g c 56 64 77 calcium, g d 0.4-0.5 1.0-1.2 1.0-1.2 vitamin A, ,ug retinol e 750 750 1 200 thiamine, mg e 0.9 1.0 1.3 riboflavin, mg e 1.3 1.4 1.8 niacin, mg e 15.2 16.5 23.8 ascorbic acid, mg f 30 50 50 vitamin D, lAg cholecalciferol f 2.5 10 10 vitamin B12 Ug f 2.0 3.0 2.5 folate, ,ug 'free" folate f 200 400 300 iron, mg f 19 ? ? a The recommended allowances are based throughout on a 'reference - woman (age 25; body weight 55 kg; weight gain during pregnancy 10 ± 2 kg; moderately active; living in a temperate climate). b FAQ (21). The daily allowance of energy for a pregnant woman is based on an additional allowance of 40 000 kcal (167.5 MJ) per pregnancy, taken over a period of about 6 months. c FAQ/WHO (23). A net protein utilization of 70 is assumed;* Protein intake considered adequate for all but a very small fraction (2.5 %) of the population.' d FAO/WHO (22). e FAO/WHO (24). The allowances are based on recommended intakes of 0.4 mg thiamine, 0.55 mg riboflavin, and 6.6 mg niacin or niacin derived from tryptophan (60 mg tryptophan = 1 mg niacin) per 1 000 kcal (4.2 MJ). f FAQ/WHO (25). The allowance of iron for a nonpregnant woman assumes that 10-25% of calories are derived from animal foods. "No figures are included . . . for recommended intakes by pregnant and lactating women, since recommended intakes must depend on knowledge of previous dietary practices." Note: Some of the figures shown above have been modified in a recent WHQ report (69a). kcal/d in milk, the provision of about 600 kcal/d (2.5 MJ) as available dietary energy should be adequate. If it can be assumed that mothers who are breast-feeding exhibit a reduced level of physical activity, the practical allowance (additional to that of a nonpregnant, nonlactating woman) may be a little lower ". Subsequently, Passmore et al., in a WHO monograph (69a), recommended an additional 550 kcal (2.3 MJ) during lactation. An FAO/WHO Expert Group (24) related the requirements of thiamine, riboflavin, and niacin to dietary energy. If the additional energy requirement imposed by lactation is less than that shown in Table 3, the allowances of those vitamins will be reduced in proportion. With reference to calcium requirements, an FAO/WHO Expert Group (22) " realized ... that, in certain populations, successful repeated pregnan- cies and lactations are achieved with a calcium intake much below the suggested allowance ". No specific comment seems to be needed with respect to the FAO/WHO recommendations on allowances of other nutrients. LACTATION AND MATERNAL HEALTH We shall not attempt to deal with aspects of breast-feeding that are difficult to quantify, such as its convenience and the satisfaction it gives to the mother. Undoubtedly breast-feeding is convenient under some social and cultural conditions and it can be a source of great pleasure to the mother as well as the baby; but these benefits may have been over- rated by enthusiasts for breast-feeding. As noted above, lactating women have increased dietary needs, and if these are not met there should be evidence of tissue depletion leading, if it continues long enough, to actual deficiency disease. Surpris- ingly, confirmatory evidence is scanty. The suspicion may be justified that concentration on the benefits of 170 HUMAN LACTATION breast-feeding for infants has led to the neglect of possible hazards for the mothers. The simplest evidence of tissue depletion is loss of body weight during lactation. Venkatachalam (92) showed in a cross-sectional study that the mean weight of Chimbu women in New Guinea fell stead- ily with increasing parity. In villages near Madras, India, Devadas & Prema (17) reported that nursing mothers on their ordinary diets lost weight but gained when food supplements were provided. On the other hand, Gopalan & Belavady (31) reported no difference, according to parity or stage of lacta- tion, in the mean body weight of poor lactating Indian women. In a prospective study of 14 such women for up to 90 weeks after delivery, " It was noticed that in some of the mothers the body weights remained practically stationary. In those that lost weight the loss ranged between 1.8 kg and 7.3 kg, and it took nearly a whole year for this loss to occur ". Gopalan & Belavady suggest that tissue losses might have been masked by increases of body water; this hypothesis does not appear to have been re-examined. In an impoverished West African rural community, Thomson et al. (89) found that lactating women maintained body weight remarkably well; in 8 women followed during 18 months of lactation, the mean weight at the end of lactation was nearly as high as it was at the beginning. It would appear, from this limited evidence, that most lactating mothers, even when living in poor nutritional circumstances, are able to maintain body weight despite raised energy requirements. The prob- lem deserves further study. Other evidence of maternal deprivation during lactation is even more limited. Atkinson & West (3) found by means of photon beam scanning that there was a 2.2% loss of calcium, on average, from the femoral shaft during 100 days of lactation in well-fed women. They suggest that depletion of calcium in bone may be normal during lactation, even in the presence of an adequate intake. Whereas this is probably true, Walker et al. (94) doubt that the losses are cumulative, even in women having had several pregnancies and lactations while taking diets with calcium contents as low as 300-400 mg daily; they found no X-ray evidence of bone depletion in Bantu women with 7 or more children, compared with similar women with 2 or fewer children. Osteo- malacia has for long been known to be precipitated by pregnancy and possibly by lactation, but such cases appear to be rather rare except among Indian immigrants to the United Kingdom (27) or severely malnourished women in oriental countries (97). It certainly does not appear to be the result of un- complicated calcium deficiency; a concurrent defi- ciency of vitamin D may be implicated. We have not traced any reports suggesting that other vitamin deficiencies occur more commonly than usual among lactating women. For reasons that are by no means clear, the literature gives the impression that, from this point of view, the hazards are greater during pregnancy than during lactation. It may be so. Hytten & Leitch (40) suggest that during pregnancy: " The fetus, using hormones as manipulators, overrides and resets the mother's homeostatic mechanisms in its own interests: it is the price of viviparity. " This power to adjust maternal metabolism is lost at parturition, and lactation does not involve the massive endocrine adjustment that characterizes pregnancy. In their study of lactating primiparae in Aberdeen, Scotland, most of whom had access to adequate diets, Hytten et al. (38) found a high incidence of " minor " complaints such as tiredness, backache, and repeated minor infections. Whether this is a feature of lactation generally or of lactation in a modern urban society where breast-feeding is not the norm remains unclear. It is, however, a matter of simple observation that lactating women in the developing countries, especially in the " backward" rural areas where breast-feeding remains an every- day, natural event, are able to undertake hard physical work; whether they, too, feel more tired than usual appears to be unrecorded. CONCLUSIONS The natural history of breast-feeding There is a large volume of published information on the prevalence and duration of breast-feeding from various parts of the world, but relatively few reports deal with large and representative samples and few give much detail. Interpretation from a nutritional point of view is often complicated by differences of terminology and by the lack of in- formation on the extent to which complementary and supplementary feeding is used. There is aston- ishingly little objective information on the be- havioural and psychological aspects of breast-feed- ing. Published information on the reasons for the rapid decline of breast-feeding in many communities seems to be based mainly on speculation and asser- tion. There is accordingly an urgent need for data on 171 A. M. THOMSON & A. E. BLACK all aspects of breast-feeding in various ecological settings. If such data are to be collected in an objective manner, the first concern of the observers must be to record the relevant facts without bias and without directly influencing the behaviour of those being observed. This is not as simple as it may seem, since in many communities the mothers themselves expect medical and nursing personnel to exhibit certain attitudes (e.g., hostility to artificial feeding or, for that matter, to breast-feeding) and may mould their responses accordingly. It may be neces- sary to secure the information indirectly, for ex- ample, during the course of some enquiry into another topic that is unlikely to cause the mother to conceal her real feelings towards breast-feeding. Thus, it would be desirable to undertake longitu- dinal studies of maternal behaviour in relation to infant-feeding in various communities and socio- economic settings, with the object of recording, without interference, what really happens and what material and psychological influences mould mater- nal behaviour. Such studies could be combined with (and, if necessary, camouflaged by) observations of the growth and health of the infants and of the physique and health of the mothers. Particular atten- tion should be paid to ways in which industrial society alters attitudes towards breast-feeding, with reference to problems of modesty, maternal activity including paid work, the influence of the media on infant-feeding methods, and " folklore " with respect to infant-feeding. Practically nothing is known about the physio- logical capacity to lactate of mothers in different ethnic groups and under different environmental conditions. Particular attention needs to be paid to the possibility that, if a mother is unable to breast- feed satisfactorily through inadequate lactation, the baby may become sick and even die, particularly in impoverished and uneducated communities where maternal breast-feeding is the rule and satisfactory substitutes are not available. This might lead to considerable selection, so that mothers who continue to lactate for long periods are those with a relatively high capacity for lactation, whereas those who stop breast-feeding early are those with a low capacity. Studies of variations in milk output and composi- tion should be attempted in impoverished communi- ties where sustained breast-feeding is usual. How is the level of milk production related to the growth, health, and survival of infants and young children? Studies should be made of the kinds and amounts of supplementary feeding in relation to adequacy of milk production. Special attention should be given to the " natural history " of breast milk production and the nutritive value of such milk after prolonged lactation, e.g., 6 months post partum. There is a need for a study of weaning practices and of the growth and health of infants and young children around the time of weaning. Is the time of weaning influenced by the health and condition of the child? In view of the observations and specula- tion of Cantrelle & Leridon (12), is it possible that the children weaned early are those whose growth and health have been failing, whereas the ones weaned late are those whose growth and health have remained favourable? Does weaning-which may be very abrupt in some communities-lead to any faltering of growth or impairment of wellbeing? How rapidly do the maternal breasts regress to the nonlactating state, and are any difficulties experi- enced? It is known that in some communities weaning marks the time when a return to sexual activity is expected. Under what circumstances and with how much consistency are such customs observed? Are there less obvious forms of parental behaviour that might create some relationship between lactation and the amount of sexual activity? Attempts might be made during studies of breast-feeding to establish how much sexual activity takes place at various stages of lactation and after weaning. Lactation and health A considerable amount of work has already been done on the anti-infective properties of human milk and on the prevalence of various infections in fully breast-fed and fully bottle-fed infants. There is, however, a need for more information, particularly from developing communities where standards of hygiene are poor. If breast-fed infants have fewer infections, is this due to a protective influence of breast milk, or to less exposure to a highly infected environment? Is the prevalence of infection related to the amount and nature of supplementary feeding? Studies of infections in infants who are wholly and partly breast-fed should be planned from a broadly ecological point of view. Information on types of morbidity other than infection in relation to infant feeding is so frag- mentary as to be little better than speculation. There is a particular need for long-term prospective studies of hypotheses that seem reasonable. Where possible, community data on infant feeding collected during studies such as those recommended above should be 172 HUMAN LACTATION used to provide baselines for long-term follow-up studies of growth and health. The low-birth-weight infant (who may be pre- maturely born or " light for dates ") presents a problem of special importance. The very small infant that is unable to suckle will quickly die in a com- munity where special facilities for feeding and care cannot be provided. Such babies are outnumbered by those that, although technically of low birth weight, are mature enough to ensure their own sur- vival under reasonable "natural " conditions. There appears to be very little information on whether such infants, when fully breast-fed, develop as well as similar infants that are fed with suitable breast milk substitutes. It is known that artificially fed babies of low birth weight tend to grow more rapidly, at least during infancy, than those who are breast-fed. Does this confer any subsequent ad- vantage from the point of view of physique, func- tion, and general health? An attempt should be made to organize long-term follow-up studies of low-birth-weight infants (both prematurely born and " light for dates ") who have been fed in different ways. Under appropriate condi- tions, randomly chosen feeding regimens might be prescribed. Observations during follow-up should include physical measurements, tests of neurological and psychological function, and records of general health. Full lactation undoubtedly causes a considerable increase in the nutritional needs of the mother. It is important to establish the extent to which these needs are being met by depletion of her own tissues. In this context, it should be remembered that preg- nancy may have been accompanied by the storage of fat and possibly of other nutrients that will be discarded after pregnancy whether or not lactation occurs. The simplest measure of maternal gain or loss is change in body weight. Where possible, observations on weight may be accompanied by metabolic and biochemical studies designed to mea- sure changes in body composition. It is also impor- tant to establish whether or not sustained lactation is accompanied by any increased incidence of clinical deficiency states. Thus, studies of maternal physique, body composition, and health during lactation are required that are related to dietary intakes and physical activity. We are well aware that few, if any, of the above- mentioned study proposals will be easy to carry out. The lactating mother is a preoccupied, busy, and often psychologically vulnerable individual-not at all the ideal subject for close observation and experi- mental " interference ". Elaborate precautions (even subterfuges: see Hytten et al. (39)) may have to be adopted to prevent or at least minimize changes induced by the observer in the very phenomena that he wishes to observe. A premium must be placed on technical adequacy: short-cut procedures and in- adequately tested techniques may do little more than add to the mass of dubiously reliable information and sheer speculation that already clutters the litera- ture. As much effort may need to be devoted to the problem of how to collect information reliably as to the collection itself. A " fact" is only as reliable as the method used to establish it. Definitions and classifications The existing literature is also less useful than it might be because terms are often poorly defined and different classifications may make accurate com- parisons impossible. There is a need for standard definitions of terms such as " fully breast-feeding ", and " weaning ". Numerical data should be ex- pressed in standardized ways that will facilitate comparisons between data from different sources; for example, the proportions of mothers who are breast-feeding exclusively might be specified at pre- scribed intervals after parturition. Finally, all reports should contain enough information to permit intel- ligent interpretation of the data from a broad eco- logical standpoint. RItSUM1t ASPECTS NUTRITIONNELS DE L'ALLAITEMENT AU SEIN Beaucoup de personnes sont fermement convaincues que l'allaitement au sein est le meilleur. Mais l'auteur, se placant a un point de vue scientifique, estime qu'il s'agit la d'une hypothese demandant i etre v6rifiee. L'allaitement au sein est une forme de comportement social tandis que la lactation est une fonction physio- logique; les donnees relatives a la frequence et A Ia duree de l'allaitement au sein n'indiquent pas necessairement que la lactation soit adequate. L'article offre un resume de rapports recents sur la frequence et la duree de 1'allaitement au sein dans divers pays. En g6neral, il est moins prolonge dans les collec- 173 174 A. M. THOMSON & A. E. BLACK tivites urbanisees prosperes et tend A diminuer dans les collectivit6s ((primitives#) soumises a 1'influence de la societe industrielle moderne. 11 faudrait etudier plus avant le phenomene. I1 semble qu'il n'existe pas d'instinct puissant poussant les femmes A allaiter. On manque d'informations fiables sur la quantite de lait secretee. D'apres la litterature, elle est de 500 A 700 ml par jour chez toutes les femmes, riches comme pauvres. Le regime suivi par la mere ne semble pas influer sur la composition du lait sauf en ce qui concerne les vitamines hydrosolubles. D'apres des calculs theo- riques, I'allaitement matemel exclusif suffit jusqu'au moment oii le poids du bebe atteint environ 5,5 kg, vers le troisieme mois. Le lait maternel a des proprietes anti-infectieuses specifiques pour le nouveau-n6 et le jeune enfant mais il est douteux que 1'enfant plus Age en beneficie encore. II ressort des documents publies sur ce sujet qu'une prolongation de l'allaitement au sein, sans amelioration de l'hygiene, n'est pas en mesure de reduire les infections. Parmi les inconvenients possibles de l'alimentation artificielle, figurent un defaut d'assimilation du calcium du a un apport elev6 d'acides gras satures et une influence sur l'apparition precoce de l'obesite. Les taux de croissance constituent un bon moyen d'evaluation de la nutrition et de la sant6 des jeunes enfants. Les taux moyens de croissance, de la naissance A trois mois, obtenus d'apres des etudes longitudinales portant sur des nourrissons entierement nourris au sein, sont satisfaisants dans les groupes riches comme dans les groupes pauvres. En general, les nouveau-nes de faible poids a la nais- sance gagnent moins au debut s'ils sont nourris au sein que s'ils sont nourris artificiellement. Mais beaucoup d'etudes se compliquent du fait que les bebes ont ete nourris au moyen d'un tube avec du lait de femme, ce qui n'est peut-etre pas comparable a l'allaitement au sein. D'apres des donnees provenant d'une collec- tivite indienne, les enfants de faible poids A la naissance, nourris au sein, etaient comparables aux bebes plus lourds. L'auteur a r6sume les recommandations de divers groupes d'experts FAO/OMS. La ration supplementaire de 1000 kcal recommandee pendant l'allaitement se fonde sur l'hypothese que le taux de conversion de l'energie alimentaire est de 60%, mais certains auteurs pensent que ce taux est superieur a 90% et recommandent une augmentation de la ration de 2,5 MJ (600 kcal), chiffre qui correspond davantage aux valeurs de I'apport energetique observees pendant I'allaitement. Le groupe d'experts FAO/OMS des besoins en calcium (1962) a admis que # dans certaines populations, des grossesses et des lactations repetees n'entrainent aucun incident, meme lorsque 1'apport de calcium est bien inferieur a la ration proposee )>. On ne dispose que de peu d'information sur les effets sur la mere d'un allaitement prolonge. II semble que la plupart des femmes puissent conserver leur poids cor- porel malgre 1'accroissement de leurs besoins ener- getiques. Certains auteurs ont emis l'hypothese qu'un deficit en calcium etait normal au cours de I'allaitement mais d'autres doutent que les pertes soient cumula- tives. Les carences vitaminiques ne semblent pas plus frequentes chez les femmes allaitantes. Les auteurs de I'article estiment que, d'une maniere generale, l'allaitement au sein associe a une lactation adequate est le meilleur. Toutefois, bien des points restent obscurs et des recherches sont necessaires dans les domaines suivants: facteurs d'ordre materiel et psychologique qui influencent le comportement quant a l'allaitement maternel; physiologie de la lactation dans les collectivites pauvres, duree de la lactation associ6e a l'aptitude d'allaiter, croissance de l'enfant et intro- duction d'aliments solides; circonstances de l'allaitement au sein, du sevrage et de la morbidite infectieuse; ali- mentation des enfants presentant un faible poids a la naissance; et effets sur la sante maternelle d'un allaite- ment prolonge. Une grande partie de la litterature existante n'est pas entierement fiable; il faudrait planifier soigneusement des travaux de recherche en mettant I'accent sur leur qualite technique. REFERENCES 1. AITKEN, F. C. & HYTTEN, F. E. Nutrition abstracts and reviews, 30: 341-371 (1960). 2. ARNEIL, G. C. Dietary study of 4 365 Scottish in- fants-1965. Edinburgh, Scottish Home and Health Department, 1967 (Scottish Health Service studies, No. 6). 3. ATKINSON, P. J. & WEST, R. R. Journal of obstetrics and gynaecology of the British Commonwealth, 77: 555-560 (1970). 4. BAILEY, K. V. Journal of tropical pediatrics and African child health, 11: 35-49 (1965). 5. BANIK, N. D. D. ET AL. Indian journal of pediatrics, 37: 95-101 (1970). 6. BECROFr, T. & BAILEY, K. V. Journal of tropical pediatrics and African child health, 11: 28-34 (1965). 7. BtHAR, M. ET AL. Archives of environmental health, 17: 814-827 (1968). 8. BLAXTER, K. L. Lactation and the growth of the young. In: Kon, S. K. & Cowie, A. T., ed. Milk: the mammary gland and its secretion, Vol. II. New York, Academic Press, 1961, pp. 305-361. 9 BOUCHALOVA, M. & OMDELKA, F. eeskoslovenskd pediatrie, 25: 545-547 (1970). HUMAN LACTATION 175 10. British medical journal, 2: 727-728 (1973). 11. BROWN, M. L. & ADELSON, S. F. Tropical and geographical medicine, 21: 53-61 (1969). 12. CANTRELLE, P. & LERIDON, H. Population studies, 25: 505-533 (1971). 13. CHUECKA, M. ET AL. Courrier, 18: 221-231 (1968). 14. CHRIsTIANIN, G. & DE GRESSI, E. Minerva pedia- trica, 22: 359-361 (1970). 15. Coy, J. F. ET AL. Medicaljournal ofAustralia, 2: 132- 134 (1970). 16. DAVIES, D. F. American heart journal, 81: 289-291 (1971). 17. DEVADAS, R. P. & PREMA, L. Journal ofnutrition and dietetics, 2: 149-153 (1965). 18. DUGDALE, A. E. British journal of nutrition, 26: 423- 432 (1971). 19. EDOZIEN, J. C. Malnutrition in Africa-need and basis for action. In : Gyorgy, P. & Kline, 0. L., ed. Malnutrition is a problem of ecology. Basle/New York, Karger, 1970, pp. 64-72. 20. FALKNER, F. ET AL. Courrier, 8: 1-11 (1958). 21. FAO Nutrition studies, No. 15: Calorie require- ments. Rome, 1957. 22. FAO/WHO EXPERT GROUP ON CALciuM REQUIRE- MENTs. Calcium requirements. Geneva, Rome, 1962 (WHO Technical report series, No. 230; FAO Nutri- tion meetings report series, No. 30). 23. FAO/WHO EXPERT GROUP ON PROTEIN REQUIRE- MENTs. Protein requirements. Geneva, Rome, 1965 (WHO Technical report series, No. 301; FAO Nutri- tion meetings report series, No. 37). 24. FAO/WHO EXPERT GROUP ON REQUIREMENTS OF VITAMIN A, THIAMiNE, RIBOFLAVINE AND NiAcIN. Requirements of vitamin A, thiamine, riboflavine and niacin. Geneva, Rome, 1967 (WHO Technical report series, No. 362; FAO Nutrition meetings report series, No. 41). 25. FAO/WHO EXPERT GROUP ON REQUIREMENTS OF AscoRBc Acm, VfTAmiN D, VITAMIN B12, FOLATE AND IRON. Requirements of ascorbic acid, vitamin D, vitamin B12, folate and iron. Geneva, Rome, 1970 (WHO Technical report series, No. 452; WHO Nutrition meetings report series, No. 41). 26. FAO/WHO EXPERT COMMIlTEE ON NUTRrTION. Eighth report. Geneva, Rome, 1971 (WHO Tech- nical report series, No. 477; FAO Nutrition meetings report series, No. 49). 27. FELTON, D. J. C. & STONE, W. D. British medical journal, 1: 1521-1522 (1966). 28. FOMON, S. J. ET AL. Acta paediatrica, Suppl. 202 (1970). 29. FOMON, S. J. ET AL. Acta paediatrica, Suppl. 223 (1971). 30. GOPALAN, C. Indian journal of medical research, 46: 317-324 (1958). 31. GOPALAN, C. & BELAVADY, T. Federation proceed- ings, 20 (1), Part III, Suppl. 7, 177-184 (1961). 32. GRANTHAM-MCGREGOR, S. M. & BACK, E. H. Ar- chives ofdisease in childhood, 45: 404-409 (1970). 33. GRULEE, C. G. & SANFORD, H. N. Journal of pediatrics, 9: 223-225 (1936). 34. GuTHmE, H. A. & GUTHRmE, G. M. Clinical pedi- atrics, 5: 481-484 (1966). 35. GYORGY, P. American journal ofclinical nutrition, 24: 970-975 (1971). 36. HARMs, L. E. & CHAN, J. C. M. American journal of diseases ofchildren, 117: 483-492 (1969). 37. HYTTEN, F. E. Neonatal studies, 4: 155-164 (1955). 38. HYTTEN, F. E. Proceedings of the Nutrition Society, 17: 57-63 (1958). 39. HYTTEN, F. E. ET AL. British medical journal, 1: 310- 315 (1958). 40. HYTTEN, F. E. & LEITCH, I. The physiology of human pregnancy, 2nd ed. Oxford, Blackwell Scien- tific Publications, 1971. 41. HYTTEN, F. E. & THOMSON, A. M. Nutrition of the lactating woman. In : Kon, S. K. & Cowie, A. T., ed. Milk: the mammary gland and its secretion, Vol. II. New York, Academic Press, 1961, pp. 346. 42. INSULL, W. ET AL. Journal ofclinical investigation, 38: 443450 (1959). 43. JAIN, A. K. ET AL. Demography, 7: 255-271 (1970). 44. JELLIFFE, D. B. ET AL. Tropical and geographical medicine, 15: 411421 1963). 45. JELLIFFE, D. B. & JELLIFFE, E. F. P., ed. Symposium: the uniqueness of human milk. American journal of clinical nutrition, 24: 968-1024 (1971). 46. JHA, S. S. Pediatric clinics of India, 4: 49-56, 57-60 (1969). 47. KAMAL, I. ET AL. American journal of obstetrics and gynecology, 105: 315-323 (1969). 48. KARMmuKAR, M. G. ET AL. Acta paediatrica, 52: 473- 480 (1963). 49. KAss, A. Tidsskrift for den norske Laegeforening, 88: 939-942 (1968). 50. KiNG, F. Lancet, 2: 335 (1972). 51. KNUTSSON, K. E. & MELLBIN, T. Journal of tropical pediatrics and African child health, 15: 4049 (1969). 52. Ko, Y. S. ET AL. Korean journal ofnutrition, 3: 87-94 (1970). 53. KRAMER, M. ET AL. Nutritio et dieta, 7: 71-79 (1965). 54. KuBAT, K. ET AL. eeskoslovenskd pediatrie, 25: 547- 550 (1970). 55. LONGO, L. D. American journal of obstetrics and gynecology, 89: 470-475 (1964). 56. MACY, I. G. & KELLY, H. J. Human milk and cow's milk in infant nutrition. In: Kon, S. K. & Cowie, A. T., ed. Milk: the mammary gland and its secre- tion, Vol. II. New York, Academic Press, 1961, pp. 265-304. 57. MASSE, G. (1969) Croissance et developpement de l'enfant a Dakar. Paris, Centre international de l'Enfance, 1969, p. 78. 58. MATA, L. J. & WYATT, R. G. American journal of clinical nutrition, 24: 976-986 (1971). 176 A. M. THOMSON & A. E. BLACK 59. MCGREGOR, I. A. ET AL. Transactions of the Royal Society of Tropical Medicine and Hygiene, 62: 341- 352 (1968). 60. McGREGOR, I. A. ET AL. Transactions of the Royal Society of Tropical Medicine and Hygiene, 64: 48-77 (1970). 61. MELLANDER, 0. ET AL. Acta paediatrica, 48, Suppl. 116 (1959). 62. MEYER, H. F. Clinicalpediatrics, 7: 708-715 (1968). 63. MIMSTRY OF HEALTH OF GREAT BRITAIN. Standards of normal weight in infancy. London, HMSO, 1959 (Reports on public health and medical subjects, No. 99). 64. MORRISON, S. D. Human milk: yield, proximate principles and inorganic constituents. Slough, Com- monwealth Agricultural Bureau, 1952 (Common- wealth Bureau of Animal Nutrition technical com- munication No. 18). 65. NELSON, W. E. ET AL., ed. Textbook of pediatrics, 9th ed. Philadelphia, W. B. Saunders, 1969, pp. 42-43. 66. NIMs, B. ET AL. American journal of diseases of children, 43: 828-844 (1932). 67. Nims, B. ET AL. American journal of diseases of children, 43: 1062-1076 (1932). 68. OBERNDORFER, L. & MEnA, W. Journal of tropical pediatrics and African child health, 14: 27-42 (1968). 69. OSBORN, G. R. Obstetrics and gynecology, 37: 126- 130 (1968). 69a. PASSMORE, R. ET AL. Handbook on human nutri- tional requirements, Geneva, 1974 (WHO Mono- graph series No. 61). 70. POPPER, K. R. Conjectures and refutations. London, New York, Basic Books, 1962, pp. 33-65. 71. PROTHERO, R. Medical officer, 121: 141-142 (1969). 72. RAJALAKSHMI, R. Tropical andgeographical medicine, 23: 117-125 (1971). 73. RAO, K. S. ET AL. Bulletin of the World Health Organization, 20: 603-639 (1959). 74. RIVIERA, J. American journal of public health, 61: 277-280 (1971). 75. ROYAL COLLEGE OF GENERAL PRACTMONERS. Practi- tioner, 209: 356-362 (1972). 76. SANJUR, D. M. ET AL. Acta paediatrica, Suppl. 200 (1970). 77. ScRImsHAw, N. S. ET AL. Interactions of nutrition and infection. Geneva, World Health Organization, 1968 (WHO Monograph series, No. 57). 78. ScRiMsHAw, N. S. ET AL. Archives of environmental health, 18: 57-62 (1969). 79. SENtCAL, J. Courrier, 9: 1-22 (1959). 80. SETH, V. & GHA1, 0. P. Indian pediatrics, 8: 452455 (1971). 81. SHUKERS, C. F. ET AL. Journal ofnutrition, 5: 127-139 (1932). 82. SOLIEN DE GONZALES, N. L. Tropical pediatrics, 62: 577-581 (1963). 83. SOLIEN DE GONZALES, N. L. Tropical andgeographical medicine, 15: 422430 (1963). 84. STANDARD, K. L. ET AL. Journal of biosocial science, 1: 153-176 (1969). 85. STAVEREN, W. A. ET AL. Journal of the American Dietetic Association, 58: 127-132 (1971). 86. SWAMINATHAN, M. C. ET AL. Indian pediatrics, 1: 255-263 (1964). 87. THAMAN, 0. P. ET AL. Indian pediatrics, 1: 428435 (1964). 88. THAMAN, 0. P. & MANCHANDA, S. S. Indian journal ofpediatrics, 35: 334-341 (1968). 89. THOMSON, A. M. ET AL. Journal of obstetrics and gynaecology of the British Commonwealth, 73: 724- 733 (1966). 90. THOMSON, A. M. ET AL. British journal of nutrition, 24: 565-572 (1970). 91. TOVERUD, K. U. ET AL. Maternal nutrition and child health: an interpretative review. Washington, DC, National Academy of Sciences, 1950, p. 129 (Bulletin of the National Research Council, No. 123). 92. VENKATACHALAM, P. S. A study of the diet, nutrition and health of the people of the Chimbu area (New Guinea Highlands). Port Moresby, Territory of Papua and New Guinea, Department of Public Health, 1962 (Monograph No. 4). 93. VENKATACHALAM, P. S. ET AL. Journal of tropical pediatrics and African child health, 13: 70-76 (1967). 94. WALKER, A. R. P. ET AL. Clinical science, 42: 189-196 (1972). 95. WHO EXPERT COMMITTEE ON MATERIAL AND CHILD HEALTH. Public health aspects of low birth weight. Geneva, 1961 (WHO Technical report series, No. 217). 96. W1DDOWSON, E. M. Nutritio et dieta, 15: 38-47 (1970). 97. WOLFE, J. J. American journal ofdiseases of children, 49: 905-911 (1935). 98. YOON, J. J. & Kim, I. D. Korean journal of nutrition, 3: 65-80 (1970). Annex RECENT REPORTS ON THE INCIDENCE AND DURATION OF BREAST-FEEDING Sample Breast-feeding (BF) behaviour Reference Europe London, England Scotland Norway Czechoslovakia Italy The Americas and the Pacific USA Rochester, USA Pennsylvania, USA Hawaii, USA Jamaica Colombia Surinam Guatemala Mexico Asia, Far East, Australasia Tasmania Korea China (Province of Taiwan) Amritsar, India Bombay, India Delhi, India Kashmir Africa Ethiopia Uganda/Sudan The Gambia Nigeria Senegal mixed population attending a child welfare clinic, 1965-68 (N = 584) national sample, 1965 (N = 3 743) 2 000 mothers of healthy term infants 393 infants in Prague 1 013 infants in Gorizia returns from 2 928 maternity hospi- tals 383 mothers, including many in the professional class 129 middle-class mothers, college community 281 infants of low-income and middle-income families in Honolulu 300 infants born in University Hos- pital, Kingston 1 412 babies of 200 mothers attend- ing clinics 202 families of various races, mostly agricultural workers Black Carib community 125 infants in a 'pre-industrial setting ' 547 mothers of children born in 1969 304 children from various areas 4 631 married mothers using in- trauterine contraceptive devices 500 children aged 1 month to 14 years admitted to hospital; poor social conditions 61 infants of street sweepers; poor social conditions 3 areas around Delhi 312 children attending hospital (as outpatients or admissions), selected at random communities of Tigre (traditional), Sidamo (more affluent), and Arussi (cattlerearing area) Acholi tribe (Nilotic), 815 rural chil- dren up to 4 years, and a small sample of urban children rural village Yoruba community in south-west Nigeria all births (about 8 500) in a rural area 41 % BF, of which 77 % for less than 3 months 85 % of infants never BF or BF for less than 1 month 96 % completely or partly BF on discharge from hospital; at 3 months, 28 % still BF 69 % BF less than 2 months 67 % wholly or partly BF for 4 months 18% wholly and 9% partly BF on discharge from hospital 41 % BF in first month 47 % BF at 2 weeks, 12 % BF more than 5 months About one-quarter BF, more in middle-income than in low-income groups 18 % wholly BF and 27 % partly BF at 3 months; 79 % wholly or mostly on bottle at 6 months average duration of full BF: 3.5 months most infants BF until 9-12 months; working mothers partly bottle feed- ing at 6 weeks to 3 months BF -started shortly after birth weaning usually complete by 2 years but often earlier 61 % fully BF at 3 months and 9 % at 6 months; partial BF until 1-1 .5 years 53 % BF initially, 17 % fully BF, and 6 % partly BF at 3 months 303 B F up to 7 months, 50 % weaned 18-23 months 94 % breast-fed; average length of lactation: 16-18 months 99 % BF initially, 89 % BF more than 6 months, 82 % BF more than 1 year; 62.5 % started weaning 6-12 months all BF initially; 64 % wholly and 26 % mainly BF 3 months; 72 % still wholly or mainly BF at 12 months almost all BF initially; at 1 year, 37 % of urban, 4.5 % of semi-urban, and 2.1 % of rural children completely weaned 96 % BF initially, 85 % for over 6 months, 71 % for over 12 months, 31 % for over 2 years in Tigre 64 % and in Sidamo 32 % of children B F for more than 1 8 months; in Arussi, 11 % BF more than 12 months BF "continued into the second year or longer" in rural areas; urban children BF 16-18 months BF "virtually universal and con- tinued for 18-24 months in most cases ' BF is continued for 2-3 years and solid food is not usually given until teeth appear children alive at 3 years: 99.5 %, 99.0 %, 96.5 %, 80.7 % BF at 6, 12, 18, and 24 months, respectively Prothero (71) Arneil (2) Kass (49) Kubht et al. (54) Christianini & De Gressi (14) Meyer (62) Harris & Chan (36) Guthrie & Guthrie (34) Brown &Adelson (11) Grantham-McGregor & Back (32) Oberndorfer & Mejia (68) Staveren et al. (85) Solien de Gonzales (83) Sanjur et al. (76) Coy et al. (15) Yoon & Kim (98) Jain et al. (43) Thaman & Manchanda (88) Jha (46) Seth & Ghai (80) Thaman et al. (87) Knutsson & Mellbin (51) Jelliffe et al. (44) McGregor et al. (60) Longo (55) Cantrelle & Leridon (12) Place