which ones the children were drawn to on any given day, because, assuming they took food from several of the bowls at each meal, they could not help but eat a diet of an excellent standard of nutrition. Davis said that her choice of food was designed to mimic the conditions of ‘primitive peoples’, though the heaping bowlfuls were surely more plentiful than any hunter-gatherer regime. The experiment proved that when your only food choices are good ones, preferences become unimportant. The ‘fifteen patterns of taste’ resulted in a single healthy whole-food diet, because of the set-up. None of the children was totally omnivorous, but nor were their likes and dislikes a problem, as they so often are in normal family life. There was no option to like unhealthy food and dislike healthy food.
Davis herself concluded that her experiment showed that the selection of food for young children should be left ‘in the hands of their elders where everyone has always known it belongs’. Instead of the ‘wisdom of the body’ Davis spoke of the ‘glaring fallibility of appetite’. It was obvious to her that there was no ‘instinct’ pointing blindly to the ‘good’ and the ‘bad’ in food. The two most popular foods overall in her study were also the sweetest: milk and fruit. Had she offered the children a free choice of ‘sugar and white flour’, those staples of a 1930s diet, it is unlikely they would have ended up in such fine fettle. Self-selection, she concluded, would have little or no value if the children were selecting from ‘inferior foods’.
The real test, Davis recognized, would be to offer newly weaned infants a choice between natural foods and processed food. This was to have been her next experiment, but the Depression dashed this prospect, as her funding ran out at the crucial moment. Davis never got the chance to test the effects on appetite of the ‘pastries, preserves, gravies, white bread, sugar and canned food’ that had in her lifetime become so popular. Davis’s experiment left a powerful legacy that took no account of the trick at the heart of it. Doctors, particularly in America, interpreted her experiment to mean that children’s appetites are inbuilt and benign, without paying attention to the way in which Davis had changed the food environment in which the babies ate. Her work was seized on as proof that our individual appetites are messages encoded with exactly the nutrients that our particular body needs. If we need protein, we will crave chicken. If we have rickets, we will naturally gorge on vitamin D until we are cured. All we have to do to eat well is listen to our cravings. Mother Nature knows best. Davis herself gave licence to such a view, commenting that the children’s successful ‘juggling and balancing’ of more than thirty essential nutrients suggested ‘the existence of some innate, automatic mechanism … of which appetite is a part’.18
Influenced by Davis, the dominant view on appetite among paediatricians became ‘the wisdom of the body’, which went along with the vogue for ‘child-centred’ learning. In 2005 Benjamin Scheindlin MD, a paediatrician, noted that Davis’s work contributed to a widespread change in attitudes in paediatric medicine from the 1930s onwards.19 Where a previous generation lamented the pickiness of children’s changeable tastes, now doctors positively welcomed childish vagaries of appetite. Dr Spock, author of the bestselling Baby and Child Care, first published in 1946, devoted ten pages to the Davis experiment. A mother, in Spock’s opinion ‘can trust an unspoiled child’s appetite to choose a wholesome diet if she serves him a reasonable variety and balance’.20 It didn’t matter if a child developed a temporary dislike of a vegetable, because their cravings would naturally give them everything they needed in the way of nutrition.
Many experts in child-rearing still think like this, operating on the assumption that children are born with special appetites for exactly the nutrients they most need and that it will all balance out, if only they are given free rein to eat what they like. A book on solving children’s eating problems that went through several reprints in the 1980s and 1990s argued that the implications of Davis’s work were that children should be given total control over food selection: let them eat cornflakes!21 As recently as 2007, a popular website about feeding children discussed Davis and concluded that there was ‘a strong biological plausibility … that children will instinctively choose a balanced diet’.22
The ‘wisdom of the body’ is an alluring thought (like maternal instinct and other biological myths). Eating would be such a simple business, if only we had little memos inside our bodies telling us what we needed to eat at each precise moment (your vitamin C levels are dropping – quick, eat a kiwifruit!). If only we liked just the stuff that was good for us and disliked anything superfluous or bad. We can certainly learn to get better at reading our body’s cues for food, but this tends to come with age and experience, as you notice little things like how pasta for lunch makes you sleepy or that a handful of nuts and a cup of Greek yoghurt keep you full for longer than white toast and jam. But children’s omnivorous bodies – after the milk stage, when breastfed infants do self-regulate – are not so wise.
Many children habitually seek out precisely the foods that are least suitable for them. They crave sugar and shun green vegetables. They neglect to drink enough water. Nutritious meals are rejected, while junk is revered. Can we really believe that a preschooler demanding a packet of the latest kids’ sugary breakfast cereal, having seen it on TV, is responding to their body’s need for certain vitamins and carbohydrate?
The scientific evidence – both from humans and rats – shows that the theory of the ‘wisdom of the body’ is flawed at best. For the theory to be true, omnivores would need to have specific appetites for the essential nutrients the body needed at any given time. This is a very unlikely proposition, given that the list of nutrients needed by omnivores comes in so many guises, depending on the environment we happen to live in. An innate appetite for the vitamin C in blackcurrants would be no use if you lived somewhere that blackcurrants do not grow. In lab conditions, rats – our fellow omnivores – have shown a very erratic ability to self-select the diet that would do them the most nutritional good. In one study, a group of rats were given a choice between a bad-tasting but protein-rich diet and a good-tasting but low-protein diet. Over the course of a week, fourteen out of eighteen rats failed to develop a preference for the food that would have done them the most good and they lost weight.23 Other trials have attempted to find out whether rats could ‘self-select’ to correct certain vitamin deficiencies; and concluded that many of them could not. With thiamine-deprived rats, the process of learning to like a thiamine-rich diet took a week or more and the rats who did not adapt quickly enough to the correct food died.24 As for human subjects, there is, notes one specialist in the field, no data to suggest innate appetites for specific foods. It does seem possible for humans to learn over time specific appetites that will correct certain imbalances – particularly a craving for salt when lacking in sodium – but that is a different matter.25
Ninety years after Davis’s experiment, the view that likes are predominantly innate – or genetic – looks shaky. When trying to get to the bottom of where tastes come from, scientists have often turned to twins. If identical twins share more food likes than non-identical twins, the chances are that there is a genetic cause. Twin studies suggest that many aspects of eating are indeed somewhat heritable. Body weight – measured as BMI – appears to be highly heritable in both boys and girls.26 So is dietary ‘restraint’, or the mysterious urge to resist eating the thing you want to eat.27 But studies that look at likes and dislikes are much less conclusive. Several twin studies have suggested that identical twins are more likely to enjoy the same protein foods than non-identical twins, but when it came to snacks, dairy and starchy foods, their likes were only marginally more similar than those of the non-identical twins.28 Overall, the evidence for tastes being heritable is very modest, accounting for only around 20 per cent – at most – of the variation in foods eaten.29
Genes are only ever part of the explanation for what we choose to eat. As one senior doctor working with obese children put it to me, you could be cursed with all the genes that make a person susceptible to heart disease and obesity and still grow up healthy, by establishing balanced food habits. ‘All of it is reversible,’ he said. Parents and children resemble each other no more in the foods they like than couples do, suggesting that nurture – who you eat with – is more powerful than nature in determining our food habits.30 Whatever our innate dispositions, our experience with food can override