I think you would find the work of Clara Davis very instructive. Here's a PDF of her remarkable paper:
www.pubmedcentral.nih.gov/picrender.fcgi?artid=537465&blobtype=pdf
Here are some excerpts for Davis' original article, my emphasis in bold. There is a lot of wisdom in her work, I think, and it lacks the preconceptions that most Americans today have about food. Some of the information (such as that hunter gatherers are healthy on their diets) has completely been eliminated from conventional wisdom.
"For every diet differed from every other diet,
fifteen different patterns of taste being presented,
and not one diet was the predominantly cereal
and milk diet with smaller supplements of fruit,
eggs and meat, that is commonly thought proper
for this age. To add to the apparent confusion,
tastes changed unpredictably from time to time,
refusing as we say "to stay put," while meals
were often combinations of foods that were
strange indeed to us, and would have been a
dietitian's nightmare-for example, a breakfast
of a pint of orange juice and liver; a supper of
several eggs, bananas and milk. They achieved
the goal, but by widely various means....
Appetite also appears to have fallibilities with
processed foods which have lost some of their
natural constituents and which have become such
important features of modern diet, e.g., sugar
and white flour. Certainly their introduction
into previously sound primitive diets has invariably
brought with it a train of nutritional evils,
and their widespread excess in civilized diets is
decried by nutritional authorities.
All the articles
on the list, except lettuce by two and spinach by
one, were tried by all, and most tried several
times, but within the first few days they began to
reach eagerly for some and to neglect others, so
that definite tastes grew under our eyes. Never
again did any child eat so many of the foods as
in the first weeks of his experimental period.
Patterns of selective appetite, then, were shown
to develop on the basis of sensory experience,
i.e., taste, smell, and doubtless the feeling of
comfort and well-being that followed eating,
which was evidenced much as in the breast-fed
infant. In short, they were developed by
sampling, which is essentially a trial and error
method. And it is this trial and error method,
this willingness to sample, that accounts for the
most glaring fallibility of appetite. From time
immemorial adults as well as children have eaten
castor oil beans, poisonous fish, toad stools and
nightshade berries with fatal results. Against
such error, only the transmission of racial experience
as knowledge can protect. Such error
affords additional proof that in omnivorous
eaters there is no "instinct" pointing blindly to
the "good" or "bad" in food.
By this time you have all doubtless perceived
that the "trick" in the experiment (if "trick" you
wish to call it) was in the food list. Confined to
natural, unprocessed and unpurified foods as it
was, and without made dishes of any sort, it reproduced
to a large extent the conditions under
which primitive peoples in many parts of the
world have been shown to have had scientifically
sound diets and excellent nutrition. Errors the
children's appetites must have made-they are
inherent in any trial and error method-but the
errors with such a food list were too trivial and
too easily compensated for to be of importance
or even to be detected.
The results of the experiment, then, leave the
selection of the foods to be made available to
young children in the hands of their elders where
everyone has always known it belongs. Even the
food list is not a magic one. Any of you with a
copy of McCollum's or H. C. Sherman's books on
nutrition and properties of foods, could make a
list quite different and equally as good. Selfselection
can have no, or but doubtful, value if
the diet must be selected from inferior foods.
Finally, by providing conditions under which
appetite could function freely and beneficently as
in animals and primitive peoples, the experiment
resolved the modern conflict between appetite
and nutritional requirements. It eliminated
anorexia and the eating problems that are the
plague of feeding by the dosage method."