In ninety-four Dutch nulliparous women the effects of a low-Na diet in pregnancy on blood pressure, energy and nutrient intake, Ca metabolism, Zn and Mg status and body composition were studied longitudinally. The women were randomly divided into an intervention group (n 41), which used a low-Na diet (mean urinary Na excretion 61 mmol/24 h) from week 14 of pregnancy until delivery and a control group (n 53; mean urinary Na excretion 142 mmol/24 h). No effect of the diet on blood pressure was observed. The use of a low-Na diet resulted in significantly reduced intakes of energy, protein, carbohydrates, fat, Ca, Zn, Mg, Fe and cholesterol. However, the women on the low-Na diet appeared to be able to adapt quite well to the reduced intake since Ca, Zn and Mg homeostasis was maintained. In the case of Ca and Mg this was probably due to the observed reduced urinary excretions of these nutrients. Non-significant reductions in weight gain (1·5 kg) and fat-mass gain (0·9 kg) over pregnancy were found in the women on the low-Na diet. No significant effects of the diet on birth weight or placental weight were observed.

The present paper presents a certain number of cultural elements which interact in the determination of the frequency of food intake. Approaches from various perspectives (historical, ethnological, anthropological, sociological) draw attention to two major aspects relating to the periodicity of food intake: the extreme cultural diversity and the continual modifications which have occurred over time and space. The various cultural models change and are subject to multiple influences, for example, cross-cultural, economic and historical. In addition, there are interactions between the models. The definitions of food intake and frequency play a major role in building up consumers' perceptions. These various perceptions are multiple (perception of self, of food and its virtues, the rules and moral values of consumption) and finally influence behaviours. Finally, and taking into account the systems of beliefs, the very nature of feeding behaviours may carry feelings of guilt for the subject. The study of real behaviours and their relationship with health is still incomplete for reasons of methodology and also of conceptual definition. In future, data collection has to take into account real behaviour as well as subjective perceptions and value judgements. A specific effort has to be made in the future to develop methodology. This should allow the collection of reliable data and particularly comparisons between studies, without oversimplifying and distorting cultural specificities.

The present study was designed to investigate the interaction between body weight and energy expenditure in well-nourished individuals. Energy expenditure was determined during a 10 d highly controlled work programme in apparently well-nourished adult male construction workers with a wide range of body weights (mean weight: 63·9 (SD 11·0, range 46·7-80·1) kg, mean BMI: 22·5 (SD 3·8, range 16·7-28·9) kg/m2). Total energy expenditure (mean: 12·68 (SE 0·73) MJ/d or 1·78 (SE 0·07) x BMR) was determined using doubly-labelled water and the energy costs of work activities by Oxylog. The energy expenditure during work (mean: 5·75 (SE 0·29) MJ/day or 3·48 (SE 0·09) x BMR) was estimated from the energy costs of individual tasks and the time spent in those tasks. The energy expenditure during discretionary time (mean: 4·37 (SE 0·58) MJ/d or 1·49 (SE 0·17) x BMR) was calculated by subtracting occupation and sleep expenditure (taken as1 x BMR) from total expenditure. Food intake and discretionary time allocation were recorded by the subjects. The energy expenditure in the programmed work activities (expressed as a multiple of BMR) showed a significant increase (P=0·035) with increasing body weight, suggesting that the assumed constancy of BMR multiples across a wide range of body weights may not be valid. This assertion was supported by theoretical calculations based on empirically derived equations. In order to avoid errors which could be interpreted as metabolic ‘adaptation’ it may be necessary to take account of body weight when using the BMR-multiple approach to estimate energy requirements at low body weights.

The present study was conducted to investigate the effect of dietary protein on urinary Ca excretion, apparent Ca absorption and Ca balance in young and elderly subjects. Young adults (n 29) and elderly persons (n 26) consumed diets containing 12% (diet A) and 21% (diet B) of total energy as protein for 3 weeks according to a randomized crossover design. Results showed no differences between the two age groups with respect to the interaction between protein intake and Ca excretion (both in urine and in faeces), apparent Ca absorption and Ca balance. Therefore analyses were done for both age groups separately and also for the whole group. In elderly persons and in the whole group the Ca excretion in faeces (as a percentage of Ca intake) was lower during the higher protein intake (elderly: diet A, 106 (SEM 7)% diet B, 86 (SEM 7) % P = 0·018; whole group: diet A, 99 (SEM 4) % diet B, 84 (SEM 4) % P = 0·003). In young adults faecal Ca excretion tended to be lower when they consumed diet B (diet A: 94 (SEM 5)% diet B: 83 (SEM 6)% P = 0·093). Relative urinary Ca excretion was greater during the higher protein intake in young adults and in the whole group while relative urinary Ca excretion was not different in the elderly (diet A: 15 (SEM 1) %, 14 (SEM 1) %, 15 (SEM 1) % diet B: 16 (SEM 1) %, 16 (SEM 1) % 17 (SEM 2) % for the whole group, the young and elderly subjects respectively, P = 0·019; P = 0·016; P = 0·243).The resulting Ca balance was not influenced by the amount of protein in the diet in young adults. Values for the elderly and for the whole group showed that the Ca balance during diet A was significantly more negative compared with Ca balance during diet B, despite the higher urinary Ca excretion during diet B. It can be concluded that increasing the protein intake from 12 to 21% of total energy ntake had no negative effect on Ca balance.

An experiment was conducted in twelve healthy middle-aged volunteers, six of each sex, with a mean BMI of 27kg/m2 to detect differences between morning and afternoon in postprandial blood glucose, insulin and C-peptide concentrations. These responses were measured following the consumption of isoenergetic meals that were high or low in fat content, at breakfast and at lunch. Over 4d each subject received the high-carbohydrate (L, 5·5 g mixed fat/meal) and moderately high-fat (M, 33 g mixed fat/meal) breakfasts and lunches, in three combinations (LL, MM, LM), or they fasted at breakfast time and received a moderately high-fat lunch (NM), in three Latin squares. Each evening a standard meal was given. Plasma glucose, insulin and C-peptide responses were greater following L than M meals and within both MM and LL treatments insulin and C-peptide responses were greater following breakfast than following lunch. The incremental C-peptide response to a fatty lunch following a fast at breakfast time (NM) was similar to that to a fatty breakfast, but the incremental insulin response for the same comparison was marginally lower at lunch (P=0·06). The relationship of C-peptide and insulin concentrations was assessed. Plasma glucose response to a fatty lunch was increased by a fatty breakfast. The relationships of these metabolic events with fat metabolism are discussed.

Urinary 5-L-oxoproline was measured during normal pregnancies in Southampton, England and Kingston, Jamaica. The CV of 5-L-oxoproline excretion in urine, determined over 7 d in a non-pregnant woman and three pregnant women, was 10–36%. Compared with non-pregnant women, urinary 5-L-oxoproline increased three to four times from early pregnancy in women in Southampton, a highly significant difference, and remained elevated at similar levels during mid and late pregnancy. For women in Kingston, the excretion of 5-L-oxoproline was similar to that of Southampton women in the non-pregnant group and during early pregnancy. However, there was a progressive increase in the excretion of 5-L-oxoproline as pregnancy advanced and by late pregnancy excretion was from three to ten times greater than the average for the non-pregnant women. There was a significant difference between the women in Southampton and the women in Kingston during mid and late pregnancy, with women in Kingston excreting twice as much 5-L-oxoproline during late pregnancy. If the excretion of 5-L-oxoproline is a measure of glycine insdciency, the results would indicate that in some pregnancies the ability of the mother to provide glycine for herself and the developing fetus is marginal or inadequate and the constraint appears more marked in Jamaica than in England.

HDL-phospholipids are determinants in reverse cholesterol transport. They are mostly derived from triacylglycerol (TG)-rich lipoproteins. Chylomicron size is important, therefore, because it is related to the ratio surface phospholipids: core TG and, thus, determines the availability of postprandial phospholipids for transfer to HDL. Eleven healthy young women each ingested four different fat loads supplemented with retinyl palmitate and containing 60 g sunflower oil (SO), oleic–sunflower oil (OSO), mixed oil (MO; (g/kg) linoleic acid 480, oleic acid 380, linolenic acid 13) or beef tallow (BT). At the peak of TG absorption for all loads (4 h) chylomicron diameters, determined by agarose-gel filtration, were larger after SO compared with OSO (P < 0·05) and BT (P = 0·06) and after MO compared with BT (P < 0·05). At 6 h chylomicron size was larger after the vegetable oils compared with BT (P < 0·05 in each case). After each fat load chylomicron size decreased at 6 and 8 h compared with that at 4 h (P < 0·05) except for OSO. Retinyl ester and TG concentrations were lower in chylomicrons after BT than after the other fats but not in the chylomicron-free serum (containing chylomicron remnants), suggesting absorption in the form of very small particles. Compared with the fasting value, the concentration of the Svedberg unit of flotation 20–400 fraction, which contains VLDL and chylomicron remnants, was lower 8 h after MO, the only fat to contain significant amounts of linolenic acid. We conclude that chylomicron size is dependent on the fatty acid composition of ingested fats and the time-course of digestion, being larger for polyunsaturated fatty acid-rich fats and in the early phase of digestion. On the basis of retinyl ester concentration there were no differences between fats in chylomicron-remnant clearance.

The use of oral contraceptive agents by women may be a factor that contributes to the observed inter-individual variability in the BMR. We, therefore, measured the BMR, body build and composition in two groups of young women and also assessed their self-reported level of physical activity. One group had been using oral contraceptive agents for a period of 6 months or more (OCA, n 24), while the other group had never used oral contraceptives (NOCA, n 22). There were no significant differences in age, body build or composition. The absolute BMR in the two groups were not significantly different when compared using an unpaired t test (OCA: 5841 (SD 471) v. NOCA: 5633 (SD 615)kJ/d). However, using an analysis of covariance, with either body weight or a combination of fat and fat free mass as covariates, the OCA group had a BMR almost 5% higher than that of the NOCA group (OCA: 5871 v. NOCA: 5601 kJ/d; P=0·002). When those subjects with high self-reported levels of physical activity were excluded, the difference in BMR between the two groups persisted (P = 0·001). An ANOVA of oral contraceptives use and phase of menstrual cycle showed significant differences in BMR with use of oral contraceptives (P=0·004) but no difference in BMR between phases of the menstrual cycle. In conclusion, the ose of oral contraceptive agents deserves consideration when conducting and analysing data from studies on energy metabolism in young women, as it results in a significantly higher BMR.

Patients with cirrhosis of the liver require an increased amount of protein to achieve N balance. However, the utilization of protein with increased protein intake, i.e. the slope from regression analysis of N balance v. intake, is highly efficient (Nielsen et al. 1995). In the present study, protein requirement and protein utilization were investigated further by measuring protein synthesis and degradation. In two separate studies, five or six patients with cirrhosis of the liver were refed on a balanced diet for an average of 2 or 4 weeks. Protein and energy intakes were doubled in both studies. Initial and final whole-body protein metabolism was measured in the fed state by primed continous [15N]glycine infusion. Refeeding caused a statistically significant increase of about 30% in protein synthesis in both studies while protein degradation was only slightly affected. The increase in protein synthesis was associated with significant increases in plasma concentrations of total amino acids (25%), leucine (58%), isoleucine (82%), valine (72%), proline (48%) and triiodothyronine (27%) while insulin, growth hormone, insulin-like growth factor (IGF)-I and IGF-binding protein-3 were not changed significantly. The results indicate that the efficient protein utilization is due to increased protein synthesis, rather than decreased protein degradation, and suggest that increases in plasma amino acids may be responsible for the increased protein synthesis. A comparison of the patients who had a normal protein requirement with the patients who had an increased protein requirement suggests that the increased protein requirement is due to a primary increase in protein degradation. It is speculated that this is due to low levels of IGF-I secondary to impaired liver function, since initial plasma concentration of IGF-I was about 25% of control values and remained low during refeeding.

The physiological effects of dietary fibres in humans are due to their physico-chemical properties. However, it is difficult to predict these effects simply by measuring certain characteristics in vitro. Studies in human subjects are still required to assess the effectiveness of new substrates. The aim of the present study in healthy human subjects was to evaluate the effects of two novel fibres, potato (PF) and maize (MF), on fasting and postprandial blood concentrations of carbohydrate and lipid metabolites as well as on stool ouput and transit time. The chemical composition, water-binding capacity (WBC) and fermentative properties of the fibres were also characterized in order to determine their possible involvement in digestive and metabolic effects. Stools, as well as breath and blood samples, were collected after consumption for 1 month of either a basal diet (control) or a basal diet supplemented with fibre (15 g/d). MF resisted fermentation better than PF and had lower digestibility. However, both fibres increased faecal output of dry matter, neutral sugars and water. There was an inverse relationship between stool weight and orofaecal transit time, although only MF significantly reduced transit time. Orocaecal transit was lengthened by PF, probably because of its high WBC. PF ingestion also decreased postprandial plasma levels of total and esterified cholesterol but had no effect on fasting concentrations. In contrast, MF lowered fasting cholesterolaemia and increased free:esterified cholesterol. These particular physiological and fermentative properties suggest that PF and MF would be suitable ingredients in a healthy diet.

Three experiments were conducted in healthy middle-aged volunteers (six males and six females in Expt 1, six males and two females in Expt 2 and twelve males in Expt 3) with a mean BMI of 27 kg/m2 to determine whether there is a difference between morning and afternoon dietary fat clearance and utilization, and to determine in what way the fat and starch contents of the meal influence postprandial blood lipid metabolites over 4·5 h. Over 4 days in Expt 1 each subject received isoenergetic, high-carbohydrate (L, 5·5 g mixed fat/meal) and moderately high-fat (M, 33 g mixed fat/meal) breakfasts and lunches, in three combinations (LL, MM, LM), or they fasted at breakfast time and received a high fat lunch (NM) in a randomized and balanced arrangement. Each evening a standard meal was given. The following effects were significant (P<0·05): plasma triacylglycerol (TAG) responses were greater following M meals; plasma TAG concentrations were greater in the afternoon than in the morning, following two meals of the same composition, although the postprandial incremental response was less following lunch than following breakfast and peak responses were reached much earlier than after breakfast; a low-fat breakfast, or fasting at breakfast time, delayed the peak TAG response to a M lunch. The plasma concentrations of non-esterified fatty acids (NEFA) and of free glycerol were higher in the afternoon following M meals at breakfast and lunch, especially in males. This response was reduced, by the L breakfast preceding the M lunch. Two M meals in succession lowered plasma HDL-cholesterol concentration. In Expt 2 each subject received a very low-fat (VL) breakfast, followed by a lunch of the same composition. Each of these meals was followed, 110 min from the start of eating, by an infusion of Intralipid 10% emulsion at the rate of 1 ml/kg body weight over 60 s. Clearance rates of Intralipid were faster in the afternoon than in the morning (P= 0·024). In Expt 3 twelve subjects were randomly allocated to either treatment MM or LM meal patterns, as given in Expt 1. These were given daily for a period of 17 d, during which the change in fasting plasma TAG concentration was similar in both treatments. On days 1, 16 and 17 responses were measured to the M lunch and to a glucose tolerance test (GTT), conducted 2 h 17 min after lunch. The post-lunch responses confirmed those found in Expt 1; but immediately following the glucose dose there was an abrupt increase in plasma TAG that was greater in treatment LM than in treatment MM (P= 0·025), whereas plasma NEFA concentration decreased rapidly in both treatments at that time (P = 0·00066)

The present study investigated the effect of monounsaturated fatty acids (MUFA) on postprandial coagulation factor VII activity. Fifteen healthy male volunteers consumed three meals containing equal amounts (40g) of fat, but providing different proportions of MUFA (12,17 and 24% energy)in random order. Fasting and postprandial blood samples were drawn every hour for 9 h. The magnitude of the postprandial triacylglycerolaemic response and the postprandial plasma nonesterifled fatty acid (NEFA) concentrations were not significantly different following the three meals. Coagulation factor VII was activated during postprandial triacylglycerolaemia but the area under the curve of postprandial coagulation factor VII activity was not significantly different following the three meals. Regression analysis showed that fasting factor VII activity was the single mast important factor affecting postprandial factor VII activity, irrespective of plasma lipid concentrations and meal fat composition. Peak postprandial factor VII activity was attained significantly earlier following the high-MUFA meal compared with the low-MUFA meal (6·33 (SD 2·16) h, 3·60 (SD 1·81)h respectively; P = 0·016). Regression analysis showed that meal MUFA content was the primary determinant of time to peak postprandial factor VII activity. Although the magnitude of postprandial coagulation factor VII activity was not affected by meal MUFA content, peak postprandial factor VII activity occurred earlier and fasting activity levels were quickly restored following the high-MUFA meal. A short-lived increase in factor VII activity may be more beneficial than a prolonged thrombotic response.

To investigate the chronological changes in conjunctival epithelium after supplementation with a massive oral dose of vitamin A, conjunctival impression cytology (CIC) with transfer was carried out repeatedly among 200 children aged 6–120 months in a randomized, double-blind, placebo-controlled study in Chandigarh (India). Significant conversion to normal CIC started 71–80 d after vitamin A supplementation and by 101–110d conversion had taken place in all children. Compared with the placebo group, plasma retinol concentration at 100d post-supplement was found to be significantly higher in the vitamin A-supplemented group P = 0·04. This study demonstrates that CIC responds to a massive oral dose of vitamin A 3–4 months after supplementation. These findings should guide future studies and evaluations in which CIC is used to assess response to vitamin A interventions.

In two experiments the potential value of diets based on casein or free amino acids (FAA) for amino acid utilization experiments were examined. In Expt 1 the optimum dietary electrolyte balance (dEB) for a casein-based diet was estimated by supplemention with 10 or 20 g NaHCO3/kg, to produce diets containing 64, 183 or 302 mmol/kg. In addition, piglet growth performance and efficiency of nutrient deposition of piglets given the casein diets were compared with two multiple protein source diets; Supercreep, a commercial multiple protein source diet or CFS (casein-fish-soyabean-sugar) or a FAA-based diet. Expt 2 was designed to compare piglet response to FAA diet stored at - 15° with twice daily feeding, with FAA diet stored at ambient temperature (13–30°) and offered ad libitum. A CFS diet was used as a positive control and the experiment was conducted over the 10–20 kg growth phase. Expt 1 used forty-eight piglets weaned at 20–22 d of age and allocated to one of six treatments formulated to contain at least 0·84 g lysine/MJ digestible energy in a randomized block design. Piglets given the CFS and Supercreep diets produced superior growth rates (518, 491 g/d) to those given a FAA diet (353 g/d) or casein diet containing 0, 10 or 20 g NaHCO3/kg respectively (365,417,390 g/d) between 5 and 20 kg live weight. Piglets given the casein and FAA diets had higher amino acid digestibilities than those given the Supercreep and CFS diets.The increase in the dEB of the casein diet from 64 to 183 mmol/kg improved piglet growth performance between 5 and 20 kg by 14%. All piglets given casein diets had similar ileal and faecal digestibilities, empty-body compositions, nutrient deposition rates and retention ratios. The results of Expt 2 showed that there was no beneficial effect on piglet performance of storing the FAA diet at -15° and feeding twice daily. Based on the results of these two experiments, neither the casein (0, 10, 20 g NaHCO3/kg) nor FAA diets were suitable for estimating amino acid utilization by the piglet. There remain unidentified factors which limit the growth performance of piglets given the casein and FAA diets.

Although homeostatic mechanisms remain of utmost importance, rhythmic changes are present also. The main ones have a period of 24 h (circadian) or about 2–3 h (ultradian). Circadian rhythms are derived from a body clock, found in the base of the brain, and from the pattern of our sleep wake cycle, including activity and meal times. These rhythms promote the regular changes between an active wake period and a recuperative sleep period. Ultradian rhythms are also widespread and reflect external (lifestyle) and internal factors. The internal factors include biochemical need and some sort of oscillator; but details of how many oscillators, and exactly where they are, remain to be established. Food intake, appetite, digestion and metabolism have been shown to illustrate these principles. Moreover, these principles become important when special circumstances exist as far as meal times are concerned; the particular diffculties of night workers is a good example.

We examined the effect of maternal chronic cold exposure, induced by winter-shearing ewes 4 weeks before their predicted lambing date, on brown adipose tissue (BAT) and liver development in lambs. Fetuses were sampled from under-fed (60% of energy requirements for maintenance and pregnancy of an unshorn ewe) shorn or unshorn ewes at 126,140 and 145 d of gestation. Lambs were sampled from ewes within 2 h of birth. Throughout gestation fetal body, BAT and liver weights were similar in shorn and unshorn groups. The level of GDP binding to mitochondrial uncoupling protein remained low throughout gestation, but increased dramatically after birth. Lambs born to shorn ewes possesd more mitochondrial protein and exhibited a significantly higher total thermogenic activity in BAT. Type I iodothyronine 5 deiodinas(EC 3.8.1.4) activity in BAT peaked at birth, as did hepatic iodothyronine Sdeiodinase activity and was significantly greater in lambs born to under-fed shorn ewes, which exhibited a higher plasma triiodothyronine concentration. Chronic maternal adaptations to prolonged cold exposure appear to enable pregnant ewes to compensate for the negative effects of under-feeding on fetal growth and development

The present work was undertaken to study the energy value of a mixture of acetic, propionic and butyric acids (0·682: 0·226: 0·092) infused intracaecally in growing pigs. A basal diet low in fibre (42 g NSP/kg DM) was given at below the requirement for maximum weight gain. In six 2-week periods, N and energy balance measurements in eight growing pigs were carried out with and without infusion of short-chain fatty acids (SCFA). Heat production was measured using open-circuit chambers and the concentration of SCFA in faeces was determined. Less than 1% of the infused SCFA was excreted in faeces illustrating the capacity of the hind-gut to absorb and metabolize SCFA. Infusion of SCFA did not affect the digestibility of nutrients and energy. However, N retention increased demonstrating that SCFA are an energy source for protein gain when pigs are fed at below the requirement of energy. Increased CH4 production together with an increased excretion of branched-chain fatty acids in faeces suggested that there was a higher microbial activity in the hind-gut during infusion. The partial utilization of the infused energyin SCFA was 0·821. A small proportion of the infused energy in SCFA was retained in protein (0·099) and a considerable amount was retained as fat (0·722).

While many studies show reduced work output in heavy labour with poor nutritional status, data among less-physically-strenuous industrial occupations is scarce. In the present study the output of ninety-two women jute-factory workers over a 1-month period was measured together with haemoglobin, BMI, arm muscle and fat area, physical work capacity (maximum O2 consumption; ), as well as socio-economic, health and activity variables. A significant correlation was found between haemoglobin and work output, which remained significant (P < 0·005) after a multipleregression analysis with other potentially confounding factors. Anaemic workers and those with values less than 1·5 1/min were significantly less active at home than non-anaemic workers (P < 0·015). Anaemic women produced an average of 5·3% less in the factory and performed an average of 6·5 h less housework per week.

Total energy intake and the frequency and size of meals are profoundly influenced by the socio-cultural context in which it occurs. Simply eating with one other person increases the average amount ingested in meals by 44% and with more people present the average meal size grows even larger. The impact of social facilitation of energy intake on the individual appears to result from genetic effects both on the individuals' sensitivity to the presence of other people and also on the number of other people an individual tends to eat with. Culture markedly affects the choice of foods in the diet and the pattern of meals over the day. However, many of the social, psychological and physical variables that influence intake are similar across cultures.

Sixty children aged 1-3 years with protein malnutrition were involved in the present study. Measurements were made of plasma transferrin receptor (TfR), haemoglobin (Hb), ferritin, transferrin and Fe in comparison with twenty apparently healthy age- and sex-matched reference children in Nigeria. Plasma TfR was measured by an immunofluorometric assay. The mean plasma concentrations of both albumin (37 g/l) and transferrin (1·88 g/l) were within reference ranges in the ‘healthy’ children. The malnourished children had severe protein deficiency as indicated by their significantly lower mean plasma albumin (24·4-28·2 g/l; P < 0·0001) and transferrin (1·24-1·53 g/l; P < 0·0001) concentrations in comparison with the reference children. In the reference children, the traditional indicators of Fe nutrition (plasma Fe, transferrin Fe saturation and Hb) were within the reference ranges, but ferritin values were raised, indicating acute or chronic infection and/or inflammation. The mean concentrations of plasma TW (4·2-5-2 mg/l) in the malnourished group were significantly lower than the mean (6·1 mgn) of the reference children (P=0·0009). In the children with severe malnutrition, none of the indicators of Fe status except Hb (81·5-86·7 g/l; P<0·0001) showed Fe deficiency, including the serum concentration of TfR and the TfR : ferritin ratio, although the Fe status was lower than in the reference children (for Fe P = 0·009; and ferritin P=0·0004). In the absence of haemodilution, the low Hb values are a clear indication that the malnourished children were Fe deficient; none of the other indices was indicative of Fe deficiency. This is the first report of TfR levels in malnourished and healthy African children.