Perhaps the most important and complex decision animals have to make once a patch has been chosen is optimal diet. Each prey item has two properties - energy value and handling time; the most profitable item would give the most energy for the least handling time (Charnov, 1976). These are preceded by search time, which is directly related to prey density. According to Charnov's Optimal Diet Theory. optimal diet is obtained in three stages:
(1) food types are ranked by their ratio of food value to handling time
(2) search time is calculated
(3) the diet begins with the food with the highest ratio, to which food types in descending order are added.
This continues for as long as the ratio of the food value to handling for each addition is greater, than the net rate of food intake without the addition. When this inequality reverses, optimal diet has been obtained (Pyke, Pulliam and Charnov, 1977). However, Shettleworth (1985) reports that pigeons appear to equate rate of pecking with amount of food, i.e. they showed a higher performance of pecking four quarter grains of food than on one whole grain (Wolfe and Kaplon, 1941); other experiments have found similar results (for example, Hall and Kling, 1960). This was the case even when it lowered the overall intake of food: they appear to behave suboptimally by over-valuing quantity in comparison with size; however, it may by that large items are more difficult to digest than small, or that the strategy is more nearly optimal in natural settings where many different sorts of items are available than in experimental situations. Optimal Diet Theory appears only to be true for simple circumstances; animals do not follow it to the letter, but rather seem to make decisions according to rules of thumb, as will be shown below.
The next decision the animal has to make is when to leave the patch he is presently in and move to a new one, as the food intake for a patch decreases with the time spent in there. Charnov (1976) has called the time when an animal leaves a patch the 'marginal value' - the time is a function of which patches the predator is visiting; the length of time between patches should be independent of the length of time the predator hunts within any one. The predator should leave the patch it is presently in when the capture rate in the patch drops to the average capture rate for the habitat - this marginal capture rate should be equalised over all patches within a habitat. Gibb (1960) proposes a 'hunting by expectation' hypothesis: an animal learns to expect a certain amount of food from each patch and leaves that patch when it has obtained that amount of food. Mellgren (1982) points out that it is assumed that the cost of searching for prey within a patch is relatively greater than the cost of travelling from one patch to another. In order to estimate its marginal capture rate, an animal has a 'giving up time' (G.U.T) (Krebs, Ryan and Charnov, 1974) - this is the length of time between the last capture and leaving the patch (based on hunting by expectation). G.U.T. is constant within an environment across patch types, and is lower in a rich environment - this has been supported by the data.
These are the main decision made by foraging animals, but other dependent choices must also be considered. Optimal patterns and speed of movement are also factors in foraging (Pyke, Pulliam and Charnov, 1977): animals tend to meander until prey is found and then increase their rate of turning, thereby remaining in the vicinity of the prey; optimal patterns result in the lowest possible frequency of path recrossing. Krebs and McCleary (1984) point out the cost of feeding: increasing the rate of the activity also increases the amount of attention that must be diverted from other matter, i.e. possibly leaving the animal vulnerable to prey. The animal must weigh such costs against the reward gained from the food. Consideration must also be paid to mate search in the long term, and nutrient constraints (particularly for herbivores, for whom plant qualities other than energy are important).
Whilst animals do obey the rules of Optimal Diet Theory, Marginal Value Theory and so on, they obviously do not carry the formulae and equations around in their heads and work them out every time they need to make an appropriate decision. They forage by what Krebs and McCleary (1984) dub 'rules of thumb', which approximate to the solutions predicted by the models. Krebs and McCleary make four general points about these rules:
(1) They are likely to be more realistic for a particular species, as they refer to the actual mechanisms used by the species in question.
(2) It does not necessarily follow that rules of thumb render behaviour suboptimal.
(3) Under some conditions, a G.U.T. rule is best, but under others it is better to stay for a fixed amount of time or a fixed number of prey (a rule of thumb).
(4) Although the rules of thumb may work well in a natural environment, they may not work so well in an experimental situation. So although it is fair to say that animals make decisions when foraging, they do not actually work them out consciously according to precise rules, but rather according to 'rules of thumb'.
One point to be made is that foraging theories usually assume that the predator is omniscient - this is obviously implausible and it is now considered that animals are actually learning all the time as they go along, combining past and present experiences. The kind of learning required is very complex, involving learning about where the food sources are, the nature of food sources, and how to obtain the food they contain (See Mellgren, 1982). Such learning is an intrinsic part of the animals' decision making.
Foraging for food is evidently not as simple as it might seem: animals continually have to make decisions, such as where to find food or what food to eat - learning whilst they do so. Many considerations have to be taken into account, including prey density, and the possibility of predators, in order for the decisions to be made. It is important, when studying foraging, not to make the mistake of believing that these decisions are conscious calculations: they are guided by 'rules of thumb' which approximate to the rules of such theories as Optimal Diet and Marginal Value, and are, perhaps surprisingly, extremely successful.
approx 1,500 words.
Bibliography.
Charnov (1976) Theoretical Population Biology, vol.9.
Gibb (1960) in Pyke, Pulliam and Charnov (1977) Quarterly Review of Biology, vol.52.
Krebs (1978) in Krebs and Davies (eds) (1984) Behavioural Ecology -
An Evolutionary Approach, Oxford.
Mellgren (1982) Journal of Experimental Analysis of Behaviour, vo.38.
Pyke, Pulliam and
Charnov (1977) Quarterly Review of Biology, vol.52.
Smith and in Pyke, Pulliam and Charnov (1977) Quarterly Review of Biology,
Dawkins (1971) vol.52.
