and manipulated within the rarefied experimental environment. This was initially done in studies with insects, using synthetic foods formulated to differ systematically in specific nutritional dimensions while holding all others constant. An analytical framework called Nutritional Geometry was invented to disentangle the individual and interactive roles of the different nutrients in dietary regulation [1,2,13,20] (Fig. 6.1).
An important question that could be solved using Nutritional Geometry is whether energy really does drive animal foraging as assumed in Optimal Foraging Theory, whether protein does, as assumed by Classical Nutritional Ecology, or whether the situation is more complex than this, and if so, how [21]. In the early experiments, the nutritional dimensions varied were dietary energy density (the ratio of macronutrients to indigestible fiber) and the ratio of the macronutrients themselves (specifically, the protein:carbohydrate ratio, because fat is not an energetic macronutrient for the herbivorous insects in the studies). Results showed that neither energy nor protein alone could explain insect foraging – in all cases, the insects distinguished the different sources of energy (protein and carbohydrate) and spread their feeding across available foods in ways that provided specific amounts and ratios of the two nutrients. When the combinations of foods provided were changed, the insects maintained the same nutritional intake by changing the relative amounts of the new foods that were eaten [22]. In one study, cockroaches were manipulated into one of three nutritional states through providing access to a single food with high, low or intermediate protein:carbohydrate ratio and thereafter allowed to mix an intake from all three foods [23] (Fig. 6.2a). The three groups initially selected very different combinations of the available foods, favoring the food that provided the nutrient of which they were previously deprived. Remarkably, when the three groups converged on a common cumulative nutrient intake (i.e. had redressed their respective nutritional imbalances), they subsequently selected similar food combinations to maintain the balance of macronutrients on which they had converged.
Figure 6.1 Basic concepts in the Nutritional Geometry framework. (a) Dietary balance. The “intake target” represents the amount and balance of nutrients that are targeted by the animal’s regulatory systems, in this case, protein vs. fat and carbohydrate combined (non‐protein energy, NPE). Foods are represented as lines, called “nutritional rails,” which originate at the origin and project into the graph at an angle determined by the ratio of the nutrients that each contains. As the animal eats, it ingests the nutrients in the same proportion as the food it is eating, and its nutritional state thus changes along the same trajectory as the rail for that food (shown by dashed arrows). The animal can therefore reach its intake target either by selecting food that has the same ratio of nutrients as the intake target (i.e. a nutritionally balanced food) (food 1) or by switching (s1 and s2) between foods that are imbalanced but nutritionally complementary (foods 2 and 3). (b) When confined to a single nutritionally imbalanced food (i.e. with a rail that doesn’t intersect the intake target), the animal is confronted by a trade‐off between over‐ingesting one nutrient and under‐ingesting the other. By feeding to the green point, it meets its target for NPE but suffers a shortage of protein of magnitude green P‐. The converse is true if the animal feeds to the blue point – it would meet its protein target but ingest an excess of NPE (blue NPE+). At the red point, it would meet its target for neither nutrient but ingest a moderate excess of NPE and moderate deficit of P. (c) Experimental protocol for testing how the animal resolves the trade‐off between over‐ and under‐ingesting nutrients when confined to imbalanced foods. Experimental groups are each assigned one of several foods differing in the ratio of the nutrients, and the shape of the resulting array of intake points reveals the regulatory strategy. Three possibilities are illustrated: the blue symbols represent prioritization of protein (i.e. feeding to the target coordinate for protein regardless of whether this involves over‐ or under‐eating NPE), the green symbols represent NPE prioritization, and the red symbols represent an intermediate response in which the regulatory systems assign an equal weighting to excesses and deficits of the two nutrients. Many other configurations are possible.
Figure 6.2 Homeostatic regulation of dietary balance by insects. (a) Self‐selection of dietary macronutrient ratios by German cockroaches (Blatella germanica). Solid diamonds represent mean + SE intakes during a 48 hours pre‐conditioning period in which the insects were restricted to food with either low, intermediate, or high protein:carbohydrate (P:C) ratio. Hollow circles represent cumulative macronutrient intakes at various intervals when subsequently allowed to self‐compose a diet from all three foods. Data from Raubenheimer and Jones [23]. (b) Benefits of nutrient balancing in female Anchomenus dorsalis beetles. Diets were experimentally manipulated to span a wide range of lipid and protein intakes, and a response surface was constructed relating these intakes to egg production (red = high and blue = low egg production). The negative diagonal is an energy isoline and the grey radials projecting from the origin are nutritional rails representing the experimental diets (as in Fig. 6.1). Egg production showed a distinct peak, varying both with the balance (across nutritional rails) and amounts (along nutritional rails) of protein and lipid eaten. Beetles allowed to compose a diet by combining the two extreme foods (labeled low P/L and high P/L), selected an intake target that corresponded with maximum egg production (the white cross, representing mean + SE intakes of the self‐selecting beetles).
Source: Modified from Jensen et al. [26].
Such studies demonstrate that insects regulate the intake of each macronutrient homeostatically; they have separate appetites for protein and carbohydrate, and these appetites interact during feeding to compose a diet with a specific mix of the nutrients. Several experiments have demonstrated that the mix insects select best meets their nutritional requirement for such functions as growth, reproduction, immunity, etc. – i.e. they self‐select a balanced diet [24–26] (Fig. 6.2b). Within the Nutritional Geometry framework, this selected point is termed an intake target. Other studies have shown that diet selection by insects changes to track specific changes in nutrient requirements. Examples include increased carbohydrate intake following prolonged flight [3], increased fat intake following depletion of energy stores during hibernation [27], and increased protein intake associated with growth and reproduction [26].
The selection of intake targets has been demonstrated in laboratory studies not only for insects but also many vertebrate species [13]. In general, animals compose diets that contain the balance of macronutrients characteristic of the foods they normally eat – carnivores select diets with a high ratio of protein to fats and carbohydrates, omnivores an intermediate ratio, and herbivores the lowest ratio. It might, at first sight, appear unsurprising, even circular, that animals select the diets that they usually eat, but in fact, it is not. Bearing in mind that these experiments are done using synthetic foods – mixtures that the species in question have never encountered in their evolutionary history, and in many cases also their lifetimes – this shows that the proximal driver of diet selection is not the foods themselves, but the nutrients they contain. In effect, it provides a window into how the nutrient‐specific appetite systems are calibrated through evolution to direct animals to eat diets that satisfy their specific nutrient needs.
Natural food environments are not always the idyllic Gardens of Eden they are sometimes assumed to be, but regularly present animals with situations where the relative availability of different