Optimal Foraging Flashcards
Choosing what to eat, how much
Cost & benefits
Optimality logic: selection will favour animals that forage most efficiently
John Maynard Smith
Cost & benefits of different behaviours e.g. maximising food intake and minimising handling time
Optimality modelling: used to determine the best course of action for an animal e.g. maximising food intake or offspring provisioning rate per unit time
Selecting what to eat: Northwestern crows always select large whelks, they drop them c. 5m onto rocks to break them open. They keep dropping until the whelk breaks.
Zach, 1979
Experiment: drop whelks from a tower
5.6 optimal height, small whelks are harder to crack
Insights from mismatch. Then a hypothesis based on cost benefit logic is found to be incorrect this can lead to further insights. The animal may not have been well ‘designed’ by selection.
The observations may have been inappropriate. An important factor may have been omitted from the model.The assumptions may not have been valid.
In a study of oystercatchers, it was found that birds selected mussels that were smaller than predicted…
… large mussels were impossible to open
Belovsky (1978)
Moose need to eat energy rich forest leaves and Sodium rich aquatic vegetation, but what is the optimal balance of these?
They have an energy constraint (minimum) which they can get from aquatic or terrestrial calories. There’s a sodium constraint meaning there’s a definite minimum intake of aquatic vegetation. And there’s a rumen constraint meaning size of rumen limits intake. Moose tend to eat the minimum sodium intake and maximise calorific intake through eating terrestrial vegetation.
Charnov’s marginal value theorem
Foraging environments tend to be patchy. Time in patch and fitness energy gain have a loading curve.
Most effective energy intake rate is tangent to the loading curve. The tangent originates from travel time.
Kacelnik 1984 How many prey items should a starling carry to its chicks?
Starlings get diminished returns as they forage because its harder to find food when carrying prey.
Assumptions of the marginal value theorem
- Travel time between patches is known
- Travel costs = patch costs
- Patch profitability is known
- No predation
- etc…
Testing MVT assumtions
To travel costs = patch costs? Great Tits, Cowie 1977. When adjusting for travel cost > patch cost, observations matched predictions.
Is patch profitability known? Downy Woodpecker (Lima, 1984). Trained to forage from logs with 24 holes - empty of with seeds. Predicted when they would stop foraging log.
No. of seeds Predicted Observed
0 or 24 1 1.7
0 or 6 6 6.3
0 or 12 3 3.5
Optimality Models and Behaviour
Provide testable quantitative predictions
Involve explicit assumptions
Illustrate the generality of decision-making
What to do when the model fails to predict observations?
Ignore it (count as acceptable error)
Accept animal is sub-optimal
Re-build model
