untitled

As a natural consequence of these multispecies effects on feeding rates, optimal

patterns of diet choice can have important implications for predator–prey inter-

actions. An adaptive predator would shift its attention away from some species

as they become more and more rare. Such behavior can have a stabilizing influence

on predator–prey dynamics, reducing the degree of variability over time of cyclical

predator–prey systems (Gleeson and Wilson 1986; Fryxell and Lundberg 1994;

Krivan 1996). This is especially likely when the growth rate of the predator is poor

on alternative prey, when the forager exhibits partial preferences, or when alterna-

tive prey do not have overlapping spatial distributions (Fryxell and Lundberg 1997).

An alternative way to model diet choice employs a technique called linear programming

to identify the optimal solution to a requirement influenced by several constraints

(Belovsky 1978; Belovsky et al. 1989). When applied to optimal foraging, this allows

researchers the means to explore more subtle hypotheses. Linear programming can

be used to predict the optimal diet for a forager which is trying to maximize its intake

of energy, while at the same ensuring that it obtains sufficient intake of a scarce

nutrient to meet its metabolic requirements. When conducted for pairwise com-

binations of alternative foods, linear programming can be readily understood from

simple graphs (Fig. 5.5).

Belovsky (1978) used linear programming to predict the optimal choice of aquatic

versus terrestrial plants by moose, based on parameters for moose living on Isle Royale,

a small island in Lake Superior. One constraint is that moose must obtain a daily

intake of 2.57 g of sodium in order to meet their metabolic requirements. Terrestrial

plants in this system are deficient in sodium, whereas aquatic plants have much higher

concentrations. Like many other herbivores, moose have limits on the amount of food

material that can be processed each day through the digestive tract. The total daily

consumption of aquatic and terrestrial plants eaten by a moose cannot exceed this

processing rate, which we call the digestive constraint. Moose also have limits on

the number of hours that can be devoted to cropping food items. Finally, each food

type has different profitability (ratio of digestible energy content to cropping time).

Thus, time spent cropping energetically poor food items (such as aquatic plants) reduces

the opportunity to look for energetically richer items (terrestrial plants). In other words,

a moose might waste valuable time eating poor food that could be spent looking for

better food.

THE ECOLOGY OF BEHAVIOR 65

5.2.3Optimal diet

selection: linear

programming

4000

3000

2000

1000

0

0 1000 2000 3000 4000 5000

Intake of terrestrial plants

Intake of aquatic plants

Minimum energy requirement

Sodium requirement

Digestive constraint

Fig. 5.5Linear

programming model of

diet selection in moose,

based on Belovsky

(1978). The lines

represent constraints

that must be met. The

range of diets that fall

within these constraints

is enclosed in the

triangle in the middle

of the diagram.

WECC05 18/08/2005 14:42 Page 65