The optimal foraging approach to diet width

Figure 9.16 The relationship between relative keel size of pitchers of Sarracenia purpurea and nitrogen added as aerial spray in plots at Molly Bog, Vermont. Dotted lines indicate 95% confidence intervals. A larger relative keel size corresponds to a reduced investment in organs of prey capture. (After Ellison & Gotelli, 2002.)

diet width and evolution

Predators and prey have undoubtedly influenced one another's evolution. This can be seen in the distasteful or

Figure 9.16 The relationship between relative keel size of pitchers of Sarracenia purpurea and nitrogen added as aerial spray in plots at Molly Bog, Vermont. Dotted lines indicate 95% confidence intervals. A larger relative keel size corresponds to a reduced investment in organs of prey capture. (After Ellison & Gotelli, 2002.)

poisonous leaves of many plants, in the spines of hedgehogs and in the camouflage coloration of many insect prey; and it can be seen in the stout ovipositors of wood wasps, the multichambered stomachs of cattle and the silent approach and sensory excellence of owls. Such specialization makes it clear, though, that no predator can possibly be capable of consuming all types of prey. Simple design constraints prevent shrews from eating owls (even though shrews are carnivores) and prevent humming-birds from eating seeds.

Even within their constraints, however, most animals consume a narrower range of food types than they are morphologically capable of consuming. In trying to understand what determines a consumer's actual diet within its wide potential range, ecologists have increasingly turned to optimal foraging theory. The aim of optimal foraging theory is to predict the foraging strategy to be expected under specified conditions. It generally makes such predictions on the basis of a number of assumptions:

1 The foraging behavior that is . .

assumptions inherent exhibited by present-day animals is in optimal foraging the one that has been favored by theory natural selection in the past but also most enhances an animal's fitness at present.

2 High fitness is achieved by a high net rate of energy intake (i.e. gross energy intake minus the energetic costs of obtaining that energy).

3 Experimental animals are observed in an environment to which their foraging behavior is suited, i.e. it is a natural environment very similar to that in which they evolved, or an experimental arena similar in essential respects to the natural environment.

These assumptions will not always be justified. First, other aspects of an organism's behavior may influence fitness more than optimal foraging does. For example, there may be such a premium on the avoidance of predators that animals forage at a place and time where the risk from predators is lower, and in consequence gather their food less efficiently than is theoretically possible (see Section 9.5.4). Second, and just as important, for many consumers (particularly herbivores and omnivores) the efficient gathering of energy may be less critical than of some other dietary constituent (e.g. nitrogen), or it may be of prime importance for the forager to consume a mixed and balanced diet. In such cases, the value of existing optimal foraging theory is limited. However, in circumstances where the energy maximization premise can be expected to apply, optimal foraging theory offers a powerful insight into the significance of the foraging 'decisions' that predators make (for reviews see Stephens & Krebs, 1986; Krebs & Kacelnik, 1991; Sih & Christensen, 2001).

Typically, optimal foraging theory makes predictions about foraging behavior based on mathematical models constructed by ecological theoreticians who are omniscient ('all knowing') as far as their model systems are concerned. The question therefore arises: is it necessary for a real forager to be equally omniscient and mathematical, if it is to adopt the appropriate, optimal strategy? The answer is 'no'. The theory simply says that if there is a forager that in some way (in any way) manages to do the right thing in the right circumstances, then this forager will be favored by natural selection; and if its abilities are inherited, these should spread, in evolutionary time, throughout the population.

Optimal foraging theory does not specify precisely how the forager should make the right decisions, and it does not require the forager to carry out the same calculations as the modeler. Later we consider another group of 'mechanistic' models (see Section 9.6.2) that attempt to show how a forager, given that it is not omniscient, might nevertheless manage to respond by 'rules of thumb' to limited environmental information and thereby exhibit a strategy that is favored by natural selection. But it is optimal foraging theory that predicts the nature of the strategy that should be so favored.

The first paper on optimal foraging theory (MacArthur & Pianka, 1966) sought to understand the determination of diet 'width' (the range of food types eaten by an animal) within a habitat. Subsequently, the model was developed into a more rigorous algebraic form, notably by Charnov (1976a). MacArthur and Pianka argued that to obtain food, any predator must expend time and energy, first in searching for its prey and then in handling it (i.e. pursuing, subduing and consuming it). Whilst searching, a predator is likely to encounter a wide variety of food items. MacArthur and Pianka therefore saw diet width as depending on the responses of predators once they had encountered prey.

Generalists pursue (and may then subdue and consume) a large proportion of the prey types they encounter; specialists continue searching except when they encounter prey of their specifically preferred type.

The 'problem' for any forager is this: if it is a specialist, then it will only pursue profitable prey items, but it may expend a great deal of time and energy searching for them. Whereas if it is a generalist, it will spend relatively little time searching, but it will pursue both more and less profitable types of prey. An optimal forager should balance the pros and cons so as to maximize its overall rate of energy intake. MacArthur and Pianka expressed the problem as follows: given that a predator already includes a certain number of profitable items in its diet, should it expand its diet (and thereby decrease its search time) by including the next most profitable item as well?

We can refer to this 'next most profitable' item as the ith item. Ei/hi is then the profitability of the item, where Et is its energy content, and ht its handling time. In addition, E/h is the average profitability of the 'present' diet (i.e. one that includes all prey types that are more profitable than i, but does not include prey type i itself), and J is the average search time for the present diet. If a predator does pursue a prey item of type i, then its expected rate of energy intake is E/h.. But if it ignores this prey item, whilst pursuing all those that are more profitable, then it can expect to search for a further J, following which its expected rate of energy intake is E/h. The total time spent in this latter case is J + h, and so the overall expected rate of energy intake is E/(J + h). The most profitable, optimal strategy for a predator will be to pursue the ith item if, and only if:

In other words, a predator should continue to add increasingly less profitable items to its diet as long as Equation 9.1 is satisfied (i.e. as long as this increases its overall rate of energy intake). This will serve to maximize its overall rate of energy intake, E/(J + h).

This optimal diet model leads to a number of predictions.

1 Predators with handling times that are typically short compared to their search times should be gener-alists, because in the short time it takes them to handle a prey item that has already been found, they can barely begin to search for another prey item. (In terms of Equation 9.1: Ei/hi is large (hj is small) for a wide range of prey types, whereas E/(J + h) is small (J is large) even for broad diets.) This prediction seems to be supported by the broad diets of many insectivorous birds feeding in trees and shrubs. Searching is always moderately time consuming, but handling the minute insects takes negligible time and is almost always successful. A bird, theoreticians are omniscient mathematicians - the foragers need not be to pursue or not pursue?

searchers should be generalists therefore, has something to gain and virtually nothing to lose by consuming an item once found, and overall profitability is maximized by a broad diet.

2 By contrast, predators with handling times that are long relative to their search times should be specialists. That is, if J is always small, then E/(s + h) is similar to E/h. Thus, maximizing E/(s + h) is much the same as maximizing E/h, which is achieved, clearly, by including only the most profitable items in the diet. For instance, lions live more or less constantly in sight of their prey so that search time is negligible; handling time, on the other hand, and particularly pursuit time, can be long (and very energy consuming). Lions consequently specialize on prey that can be pursued most profitably: the immature, the lame and the old.

3 Other things being equal, a predator should have a broader diet in an unproductive environment (where prey items are relatively rare and s is relatively large) than in a productive environment (where s is smaller). This prediction is broadly supported by the two examples shown in Figure 9.17: in experimental arenas, both bluegill sunfish (Lepomis macrochirus) and great tits (Parus major) had more specialized diets when prey density was higher. A related result has been reported from predators in their natural setting - brown and black bears (Ursos arctos and

U. americanus) feeding on salmon in Bristol Bay in Alaska. When salmon availability was high, bears consumed less biomass per captured fish, targeting energy-rich fish (those that had not spawned) or energy-rich body parts (eggs in females, brain in males). In essence their diet became more specialized when prey were abundant (Gende et al., 2001).

4 Equation 9.1 depends on the profitability of the ith item (Ei/hi), depends on the profitabilities of the items already in the diet (E/h) and depends on the search times for items already in the diet (s) and thus on their abundance. But it does not depend on the search time for the ith item, si. In other words, predators should ignore insufficiently profitable food types irrespective of their abundance. Re-examining the examples in Figure 9.17, we can see that these both refer to cases in which the optimal diet model does indeed predict that the least profitable items should be ignored completely. The foraging behavior was very similar to this prediction, but in both cases the animals consistently took slightly more than expected of the less profitable food types. In fact, this sort of discrepancy has been uncovered repeatedly, and there are a number of reasons why it may occur, which can be summarized crudely by noting that the animals are not omniscient. The optimal diet model, however, does not predict a perfect correspondence between observation and expectation. It predicts the sort of strategy that will be favored by natural selection, and says that the animals that come closest to this

(a) Bluegill sunfish

Low density Medium density High density 0 0.4 0.8 0 0.4 0.8 0 0.4 0.8

Ratio encountered

Prediction of optimal diet theory

Observed ratio in diet

Low density 0 0.4 0.8

Proportion encountered

Predicted proportion in diet Observed proportion in diet

Large prey Medium prey Small prey

High density I High density II High density III 0 0.4 0.8 0 0.4 0.8 0 0.4 0.8

Large prey Small prey

handlers should be specialists specialization should be greater in productive environments the abundance of unprofitable prey types is irrelevant

Figure 9.17 Two studies of optimal diet choice that show a clear but limited correspondence with the predictions of Charnov's (1976a) optimal diet model. Diets are more specialized at high prey densities; but more low profitability items are included than predicted by the theory.

(a) Bluegill sunfish preying on different size classes of Daphnia: the histograms show ratios of encounter rates with each size class at three different densities, together with the predicted and observed ratios in the diet. (After Werner & Hall, 1974.)

(b) Great tits preying on large and small pieces of mealworm. (After Krebs et al., 1977.) The histograms in this case refer to the proportions of the two types of item taken. (After Krebs, 1978.)

strategy will be most favored. From this point of view, the correspondence between data and theory in Figure 9.17 seems much more satisfactory. Sih and Christensen (2001) reviewed 134 studies of optimal diet theory, focusing on the question of what factors might explain the ability of the theory to correctly predict diets. Contrary to their a priori prediction, forager groups (invertebrate versus ectothermic vertebrate versus endothermic vertebrate) did not differ in the likelihood of corroborating the theory. Their major conclusion was that while optimal diet theory generally works well for foragers that feed on immobile or relatively immobile prey (leaves, seeds, mealworms, zooplankton relative to fish), it often fails to predict diets of foragers that attack mobile prey (small mammals, fish, zooplankton relative to insect predators). This may be because variations among mobile prey in vulnerability (encounter rate and capture success) are often more important in determining predator diets than are variations in the active choices of predators (Sih & Christensen, 2001). 5 Equation 9.1 also provides a context for understanding the narrow specialization of predators that live in intimate association with their prey, especially where an individual predator is linked to an individual prey (e.g. many parasitoids and parasitic herbivores - and many parasites (see Chapter 12)). Since their whole lifestyle and life cycle are finely tuned to those of their prey (or host), handling time (M) is low; but this precludes their being finely tuned to other prey species, for which, therefore, handling time is very high. Equation 9.1 will thus only apply within the specialist group, but not to any food item outside it.

On the other hand, polyphagy has definite advantages. Search costs (O) are typically low - food is easy to find - and an individual is unlikely to starve because of fluctuations in the abundance of one type of food. In addition, polyphagous consumers can, of course, construct a balanced diet, and maintain this balance by varying preferences to suit altered circumstances, and can avoid consuming large quantities of a toxin produced by one of its food types. These are considerations ignored by Equation 9.1.

Overall, then, evolution may broaden or restrict diets. Where prey exert evolutionary pressures demanding specialized morphological or physiological responses from the consumer, restriction is often taken to extremes. But where consumers feed on items that are individually inaccessible or unpredictable or lacking in certain nutrients, the diet often remains broad. An appealing and much-discussed idea is that particular pairs of predator and prey species have not only evolved but have coevolved. In other words, there has been an evolutionary 'arms race', whereby each improvement in predatory ability has been followed by an improvement in the prey's ability to avoid or resist the predator, which has been followed by a further improvement in predatory ability, and so on. This may itself be accompanied, on a long-term, evolutionary timescale, by speciation, so that, for example, related species of butterfly are associated with related species of plants - all the species of the Heliconiini feed on members of the Passifloracaea (Ehrlich & Raven, 1964; Futuyma & May, 1992). To the extent that coevolution occurs, it may certainly be an additional force in favor of diet restriction. At present, however, hard evidence for predator-prey or plant-herbivore coevolution is proving difficult to come by (Futuyma & Slatkin, 1983; Futuyma & May, 1992).

There may seem, at first sight, to be a contradiction between the predictions of the optimal diet model and switching. In the latter, a consumer switches from one prey type to another as their relative densities change. But the optimal diet model suggests that the more profitable prey type should always be taken, irrespective of its density or the density of any alternative. Switching is presumed to occur, however, in circumstances to which the optimal diet model does not strictly apply. Specifically, switching often occurs when the different prey types occupy different microhabitats, whereas the optimal diet model predicts behavior within a microhabitat. Moreover, most other cases of switching involve a change in the profitabilities of items of prey as their density changes, whereas in the optimal diet model these are constants. Indeed, in cases of switching, the more abundant prey type is the more profitable, and in such a case the optimal diet model predicts specialization on whichever prey type is more profitable (that is, whichever is more abundant; in other words, switching).

Was this article helpful?

0 0
Lawn Care

Lawn Care

The Secret of A Great Lawn Without Needing a Professional You Can Do It And I Can Show You How! A Great Looking Lawn Doesnt Have To Cost Hundreds Of Dollars Or Require The Use Of A Professional Lawn Care Service. All You Need Is This Incredible Book!

Get My Free Ebook


  • odovacar
    What does optimal foraging theory predict about an animal s foraging behavior?
    3 years ago

Post a comment