THE SELF-SELECTION OF DIETS 



Di'RiiXG the last three decades workers in agricidtural stations and in nutri- 

 ' tion laboratories have made sporadic attempts to test the ability of ani- 

 mals to select their own diets. Eward^" reported the first observations on farm 

 animals. He found that hogs which had access to a number of different food- 

 stuffs—whole meal, meat meal (60 per cent protein), whole oats, linseed-oil 

 meal, wheat middlings, charcoal, limestone, salt, water, all offered in separate 

 containers— showed excellent growth. One hog was the largest that had been 

 raised at the Iowa Agricultural Station. Pearl and Fairchild^ reported that 

 chickens which had access to a variety of natural foods grew better than those 

 that were given the regular diet. Dove^ also found that chickens thrived on 

 self-selection diets. Lewis," and Stearns and Hollander" successfidly used "cafe- 

 teria" feeding systems for pigeons. The latter authors found that choices made 

 from corn, peas, wheat, and kaffir corn, offered in separate compartments, 

 greatly speeded up egg production and spillage and wastage of food were re- 

 duced. Godden" found that animals grazing on uncultivated pastures selected 

 grass from parts of the pastures which had a higher mineral content. Orr^ has 

 reported numerous instances of self-selection of diets in cattle. One of the 

 most common observations is that animals that graze in pastures with a low 

 phosphorus content manifest a craving for foods with a high phosphorus 

 content. The observations made by Theiler, Green and Viljoen" on cattle in 

 South Africa constitute a classical example of reactions of animals to a phos- 

 phorus-deficient diet. Osborn and Mender" made the first self-selection studies 

 on rats under laboratory conditions. They found that when rats were offered 

 pairs of diets, one inferior, the other superior, the rats regularly selected the 

 latter. Mitchell and Mendel" later made similar observations. Studies have 

 also been made on human beings, most notably by Davis,^'^^ who foimd that 

 very young children given a free choice of a great variety of natural foods 

 grew and thrived. Sometimes for several days a child ate scarcely anything but 

 eggs, or large amounts of fats, and the intake of the different foods fluctuated 

 markedly from day to day, or week to week. Sweet" has also reported favorable 

 results of free-choice methods of feeding children. 



In most instances workers interested in self-selection studies have based 

 their studies on the belief that animals or human beings could make beneficial 

 selections only from so-called natural foods, that is, foods which were natural 

 to animals of a given species. More recent experiments, in which it was shown 

 that rats are able to make beneficial selections from purified substances, have 

 opened up a new attack on the problems concerned with the self-selection of 

 diets. From these studies, made chiefly on rats, we have learned the following 

 principles: 



i.Nor7nal rats make beneficial selections from purified substances: carbo- 

 hydrates, fats, proteins, minerals, vitamins. Rats have been kept several 

 hundred days on a choice of sixteen purified substances offered in separate 



