Foraging Flashcards
Finding food and the Search Image
How do animals determine what a particular type of food looks (and smells, and feels) like?
One way animals determine what sort of food to search for is through the development of a search image theory.
First proposed by Luke Tinbergen in 1960, the idea behind a search image is that when animals encounter a prey type more and more, they form a representation of that target—the prey—and this representation or image becomes more and more detailed with experience, so that the forager becomes more successful at finding that type of prey (L. Tinbergen, 1960; White and Gowan, 2014).
Some researchers argue that foragers are keying in on one or two salient attributes of the prey (color, movement, pattern), while others argue that the search image formed is closer to some sort of representation of the entire prey item (Langley, 1996; Pietrewicz and
Kamil, 1979; Reid and Shettleworth, 1992; Van Leeuwen and Jansen, 2010). In either case, animals are learning something relevant about their prey, and most animal behaviorists think that the use of the search image likely evolved as a response to the difficulty of finding cryptic prey, and assessing what is prey and what is not (Balda et al., 1998; Kamil and Balda, 1990; Zentall, 2005).
Optimal Foraging Theory
How does foraging theory predict where animals will forage and what they will eat?
Optimal foraging theory is a class of models that predicts what and where animals should eat, the effect of hunger state on risk-sensitive foraging, and much more
foraging means searching for and consuming food.
There’s a famous dilemma in economics known as the traveling salesman’s problem. If a salesman has to visit many cities on her route, and visit each one only once, what is the optimal path to take?
Foraging animals face a dilemma. Suppose an animal has information on a series of unique and separate prey items in its environment: what is the optimal path to obtain as many of these items as possible? Are
animals capable of fine-tuning the foraging strategies that lead them on such paths? Fujioka and colleagues set out to answer these questions by studying bat foraging behavior (Fujioka et al., 2016).
Fujioka’s team simplified the question to this: suppose that two prey items are available to a foraging bat. Does the bat fly in a manner that is best suited to capture the first of these items, or do its foraging sorties take into account the best path to obtain both items? The researchers began by building a mathematical model that predicted what path a bat would take in each of these cases. They then allowed bats to forage in a large outdoor arena that had 32 microphones stationed around it, so that the echolocation clicks bats use to home in on prey could be recorded.
What Fujioka et al. found was that the bats tended to fly along paths that took into account how best to obtain both food items, not just the first item they targeted. This strategy seems to pay off, as bats that flew along a trajectory that was best suited to capture the first
prey item, not both prey items, rarely were successful in obtaining that second prey item, while bats that foraged more like a traveling salesman planning ahead often obtained both prey items.
The mathematical models within the theory often employ a mathematical technique known as optimization theory to predict various aspects of
animal foraging behavior within a given set of constraints. predictions tested by experiments and observation.
Although there are many optimal foraging models in the animal behavior literature, we will examine three models, which address the following questions:
1) What food items should a forager eat?
Optimal foraging models of prey choice predict that the encounter rate with the less profitable of two prey items does not affect whether that item should be added to the diet or not. Instead, a critical encounter rate with the most profitable item is what determines which prey items should be added to a diet.
For example, A female cheetah (the forager) has killed a hare (the prey). In making the decision whether to take hares rather than some other prey into the diet, models of animal foraging behavior assume that foragers will compare the energy value (the amount of calories provided to the forager by consuming this prey), encounter rate (how often the prey is encountered by the forager), handling time (time for the forager to kill and ingest the prey) for each putative prey, and profitability (defined as energy value divided by handling time). (Based on Caro, 1994a).
We will assume that the prey type with the highest profitability denoted as prey type 1—is always taken by a forager. If we only have two prey items, this means that the optimal diet problem boils down to two questions—should prey type 2 also be taken, and if it should, under what conditions? We begin by examining the assumptions that are made in the basic optimal prey choice model (Charnov and Orians, 1973; Krebs and McCleery, 1984).
The model assumes:
• Energy intake from prey can be measured in some standard currency (for example,
calories).
• Foragers can’t simultaneously handle one prey item and search for another.
• Prey are recognized instantly and accurately.
• Prey are encountered sequentially.
• Natural selection favors foragers that maximize their rate of energy intake.
A bit of mathematical analysis produces a counterintuitive, prediction. Recall that we are asking whether a forager that consumes prey 1 should add prey type 2 to its diet. The decision of whether to add prey type 2 to the diet is not dependent on the forager’s encounter rate with prey type 2; rather, it is related to its encounter rate with prey type 1. This basic prediction of optimal diet choice has been tested many times.
Optimal prey choice model in great tits was examined by Krebs and his colleagues in four density conditions. With a knowledge of encounter rates, handling times, and energy values, the researchers were able to predict the birds’ optimal diet of larger, more profitable and smaller, less profitable prey. (Based on Krebs, 1978, p. 31).
looked at prey item selection and size given the choice of 2 worms (large and small). big worms provide more energy and has similar handling time to the small worm, thus bigger worms are more profitable. the optimality modelling suggests that the encounter rate with smaller prey does not influence the consumption of smaller prey. it is instead solely dependent on encounter rate with larger prey. how often you pick a small worm depends on how often you find larger ones. at low prey density where large and small prey items are encountered equally, the birds are not choosy and take both items on offer.
2) How long should a forager stay in a certain food patch?
The marginal value theorem predicts how long a forager should stay in a given patch, given that it can leave and travel to feed in less depleted patches.
For example, how long should a hummingbird spend sucking nectar from one flower, given that there are other flowers available? Eric Charnov developed an optimality model to answer this (Charnov, 1976; G. Parker and Stuart, 1976). in order to get to these patches (defined as a clump of food that can be
depleted by a forager), the forager must pay some cost—energy associated with travel, time lost while traveling, increased rate of predation, and so forth—associated with traveling between patches. Assuming that we know the rate of food intake, Charnov’s marginal value theorem (a compromise between costs of moving between patches, declining rate in intake, and promise of richer pickings elsewhere) makes a series of testable predictions regarding patch residence time. Cowie calculated the optimal time to stay in a patch as a function of travel time between patches. The amount of time birds spent in a patch matched the optimal time predicted by the marginal value theorem
obtaining specific nutrients?
energy value is not the only consideration. e.g. sodium (salt) is a much needed resource, lost in urine, scarce, and can produce “specific hunger”.
study by Gary Balovksi in Canada with a Canadian moose, like most large animals they need salt. they acquire it primarily from aquatic plants which are relatively energy poor. while taking time to look for salt rich, energy poor plants, this takes away precious time to look for energy rich salt poor plants they might find on land.
He produced a linear programming model which charts the various constraints on the amount of terrestrial food and aquatic food that a moose consumes on a daily dietry intake. Balovsky worked out that the first constraint is that moose need at least 750 grams of aquatic food each day to satisfy the sodium they need. the second constraint is that moose must get enough energy (3kilos of aquatic/2kilos terrestrial food). the third constraint is the size of the stomach of the moose (gut capacity). this is the range of possible diets the moose could eat.
from actually implementing this model, he observed the dietry balance of a moose between terrestrial and aquatic foods, and found that moose eat actually enough food to satisfy their sodium need and maximising the amount of energy that they obtain from their diet. called them energy maximisers, subject to other constraints that face them.
3) How does variance in food supply affect a forager’s decision about what food types to eat?
Theory predicts that hungry foragers will be risk prone—willing to assume greater variance (risk) in food intake than less hungry individuals.
The mean number of food items that the forager can
expect in both patches is identical (eight), but the variance (risk) in food intake is greater in patch 2. Should our forager take the differences in the variability into account when deciding between patches?
Hunger state and how it relates to the value (utility: Bernoulli, 1738) associated with food items plays a key role in many risk-sensitive foraging models. There are three basic value or utility functions that a forger can have:
1) A linear utility function, where every additional food item is valued equally.
2) A convex utility function, where every additional food item has less value. This sort of utility function would be one that we might expect a fairly satiated forager to
have. (Think about the value to you of a slice of cake after you have just had two ice cream cones. It might be worth something, but not what it would have been
worth before consuming the ice cream.)
3) A concave utility function, where every additional food item is worth more and more (to a limit), as might be the case for a very hungry forager
Because of a mathematical theorem called Jensen’s inequality (Jensen, 1906), fairly satiated foragers have a convex utility function and they are predicted to be risk
averse—they should prefer to forage in patches with low variance. Very hungry foragers, with their concave utility function, should be risk prone, and prefer high variance patches. Foragers with a linear utility function should be indifferent to foraging related variance.
the outcome in our risky patch:
- it is worth taking the chance of getting no food at all.
lecture material
Like all animals, great tits (Parus Major) are never more than a few missed meals (16g) from death by starvation. Finding food is the number one priority, a life-or-death matter. small food items to eat, have small fat deposits and high metabolic rate, so meals need to be frequent, little but often.
however anacondas can go months between feeds depending on the size of the prey item (227kg) they have eaten, partly because it takes a long time to digest and have a slow metabolic rate while doing so.
Growing Food
some species defend feeding territories as a way of ensuring access to food. Like Sheldon and his spot on the sofa, territory owners are invested in their territory and can benefit from accumulated knowledge about it, being able to predict where to find food efficiently. other species provide a stable food supply by hiding caches of food for future use.
territoriality: defending food
defending an area and its resources is quite common;
- benefits - ensures near-exclusive access to food supply enables more efficient use of patches. e.g. don’t return to a patch until depleted resource has been replenished.
- costs - takes time and energy to defend, risk of injury
Study of Pied Wagtails by Nick Davies. In winter wagtails set up territories along stretches across the river bank, and defend these, and these territories give a renuable supply of insects that wash up on the banks of the river. so owners forage efficiently, usually it takes about 40mins to try to patrol the entire stretch of water lying in their territories. in the course of those minutes, by the time they come back to a particular spot, food may now be available that wasn’t 40mis ago. occasionally however, owners may tolerate satellites (another individual) to share their territory, therefore the time taken to patrol the area reduces to 20mins which comes with costs and benefits.
- benefits - help with defence, satellites lack alternatives
- costs reduced food availability or predictability.
when should we expect this?
owners tolerate satellites only when food density is high. this is a case of conditional cooperation dependent on costs/benefits.
———————————————————————————
second food management strategy - caching
this is storing food for periods of scarcity (winter) must be hidden to avoid robbing. introduces selection pressure for memory of cache sites.
we can make some predictions on caching behaviour:
1) we should expect higher selection for (spatial) memory in species that cache…
there is a positive correlation between the extent of caching behaviour and the hippocampal size relative to overall body mass. Jays and Magpies are good stores and have bigger hippocampuses compared to Alpine Chough’s and Jackdaw’s. so caching behaviour has produced selection pressure for brain areas relevant for memory.
2) within a cache species, we should expect a stronger selection for spatial memory within populations that live in harsher environments.
In the study of inter-population variation for caching, birds were captured from Alaska (food poor) and Colorado (food rich). In aviaries, provided with places to cache sunflower seeds. birds eat the same amount of food dependent on which population they come from, however individual birds from different populations cache food. those taken from Alaska were much more likely to cache food than eat it than those from colorado. Alaskian birds were more efficient at locating their caching sites than were Colorado birds (fewer errors). Although Alaskan birds were smaller, their hippocampal volume was larger and contained more neurons.
Caching and Future Planning
Western Scrub Jay (type of bird) can not only cache, but plan to cache in particular locations in anticipation of future needs.
Raby et al. (2007) study investigated these birds by confining them to a cage and fed only using ground pine nuts. they are grounded therefore cannot be cached, and have to rely on eating the powdered nuts where and when they could get it.
1st experiment - over 6 days birds were trained to expect one of 2 components in early morning, get nuts from blue tray or get an empty pink tray. later evening on day 6, given unexpected access to bowl of seeds, along with a caching tray added to the compartments. so they have the choice, should I cache my food in the blue tray or the pink tray. should not put it in blue tray as you would have food to eat anyway, so you put it in the empty tray so you have food whenever you are given it.
2nd experiment - then trained to associate 2 components with different food, a blue tray with peanuts, and a pink tray with dog kibble (both enjoy). if on the sixth day they are given peanuts to store, they should store them in the pink tray, and if they are given kibble they should store them on the blue tray. because that means on the morning of day 7, they would be able to engage in their varied diet (eat both foods). and that’s what they do.
———————————————————————————
ants are one of the few species on the planet that
grow their own food. Currie and his colleagues proposed that not only have ants evolved a complex relationship with their food source (fungi), but a means to protect their food source from destruction has
also evolved in this system (Cafaro and Currie, 2005; Currie, Mueller, et al., 1999; Currie, Scott, et al., 1999; Figure 11.15).
Four lines of evidence support this hypothesis.
- First, all twenty species of the fungus-growing ants Currie and his team examined had Streptomyces (produces many antibiotics that kill
other bacteria) bacteria associated with them.
- Second, ants transmit the Streptomyces across generations, as parents pass the bacteria on to
offspring.
- Third, when male and female reproductive ants are examined (before their mating flights), only females possess Streptomyces. This is critical, as only females start new nests that will rely on the Streptomyces to produce antibiotics, and only females are involved in
“cultivating” fungus gardens.
- Fourth, and most important, the bacteria found on fungus-growing ants produce antibiotics that wipe out only certain parasitic diseases. When Currie’s team tested the antibiotics produced by Streptomyces, they found that they were effective only against Escovopsis, a serious parasitic threat to the ants’ fungus garden.
How do group social dynamics affect foraging?
Compared with solitary foraging, foraging while in groups often, but not always, increases the foraging success of individuals. Benefits of group life include the
ability to learn about various aspects of the foraging environment from observing others.
co-operative hunting is the first benefit to foraging in groups also seen in Bottlenose dolphins in Florida Bay and chimpanzees in the Thai Forest.
2nd benefit of foraging in groups relates to finding prey. non-coordinated benefits can increase with foraging group size. an example of this in Bluegill Sunfish seen above.
3rd benefit in foraging in groups is public information, using other’s behaviour to monitor changing environmental conditions.
The dynamics of group life can affect both how animals forage and foraging success. In this section we explore the following questions:
1) How does group size affect animal foraging behavior?
increasing the number of foragers in a group can increase the amount of food each forager receives (Krause and Ruxton, 2002). This may occur because more foragers flush out more prey, or because cooperative hunting creates a division of labor between different group members and increases the success rate of the cooperator.
Bluegill sunfish feed on small aquatic insects that live in dense vegetation. These items need to be flushed from substrate, and increasing the size of foraging groups may flush more prey out of the vegetation (Bertram, 1978; Mock, 1980; Morse, 1970). Mittlebach (1984) placed 300 amphipods (food) in a large tank, and found a positive relationship between foraging group size and individual foraging success up to a group size of four fish (Figure 11.16). The increased feeding rate per
individual was due to two factors. First, more prey were flushed when group size rose. Second, prey clumped together, so when one group member found amphipods, others swam over to this area and then
often found food themselves.
2) What role does cooperation within groups play in foraging?
In Christophe and Hedwige Boesch’s comparison of
hunting patterns across Gombe and Tai chimp populations, they uncovered differences in hunting strategies between these populations. In Tai chimps, hunting success was positively correlated with group
size in a nonadditive fashion: with each new hunter, all group members receive more additional food than they did when the last new hunter was added to the group (up to a limit). In addition to these group-size effects, Christophe Boesch found evidence of cooperation in Tai chimp hunting behavior (Boesch, 1994a; Gomes and Boesch, 2009, 2011). Very complex, subtle, social rules exist that regulate access to fresh kills and assure hunters who cooperated greater foraging success than those that fail to join a hunt.
3) How does work on the public information available in some groups shed light on animal foraging?
One way for a forager to assess a whole host of environmental variables—food availability, predators, and so on—is to use public information, that is, information based on the actions of others, as a cue to changes in environmental conditions (Beauchamp et al., 2012; Danchin et al., 2004; Valone, 1989; Valone and Templeton, 2002). social foragers can use the failed foraging attempts of their groupmates as additional information about when they themselves should leave a patch of food.
Jennifer Templeton and Luc-Alain Giraldeau tested this
prediction in starlings (Sturnus vulgaris). 30 cups, many empty, some containing food. target birds (bird we are focusing on) were fed alone, or with another bird that had previously been given chance to sample (few/all) cups. Two results support the predictions of public information models.
First, when tested on completely empty feeding patches, B1 (target) birds left such patches earlier when paired with the B2 (partner) bird had complete information about patch than when foraging alone. demonstrates that the target bird is sensitive to information from the partner bird.
Second, B1 birds left patches earliest of all when paired with B2 birds that had complete information about the patches (as compared with those B2 birds with only partial information). social learning is crucial; using other’s behavior to learn about novel food sources or ways to obtain it.
The ability to learn relies on neural capacity and behavioural plasticity - need relatively large forebrains. Lefebvre et al. (1997): compared forebrain size and foraging innovation in birds. in this study, foraging innovation included either ingesting new food type or using a new foraging technique. the study gathered 322 different foraging innovations.
what these researchers found in birds from north America and the British Aisles was that forebrain size (relative to brain stem) was positively correlated with frequency of foraging innovations. it also predicts ability to survive in novel environment: relative brain size predicts invasion success in birds introduced to new environments and foraging innovation predicts invasion success. Sol et al. (2005).
Natural Selection, Phylogeny, and Seed Caching
Some birds and mammals can remember where they have stored thousands of different food items (Raby and Clayton, 2010; Gibson and Kamil, 2009; Smulders et al., 2010).
How is that possible? How can an animal find scores of food items, often hidden, that are scattered across its environment over the course of months?
HIPPOCAMPAL SIZE AND CACHING ABILITY
Sue Healey and John Krebs examined memory, hippocampal volume, and food caching in seven species of corvid birds in which not only does food caching play an important role, but the location of 6,000 to 11,000 seeds must be remembered for nine months. (Healey and Krebs, 1992).
The volume of the hippocampal region relative to body
mass was positively correlated with the extent of food storing in six species of birds: (A) alpine chough, (B) jackdaw, (C) rook and crow combined, (D) red-billed blue magpie, (E) magpie, and (F) European jay.
A PHYLOGENETIC APPROACH TO STUDYING CACHING (storing food) ABILITY
PHYLOGENETIC definition: relating to the evolutionary development and diversification of a species or group of organisms, or of a particular feature of an organism.
To better understand the evolutionary history of caching behavior in corvids, Selvino de Kort and Nicky Clayton constructed a phylogeny of forty-six species from this family of birds (de Kort and Clayton, 2006).
Each species was then categorized on its tendency to cache seeds.
Based on a combination of published laboratory and field studies, each of the species was placed into one of three categories.
- Non-cachers were defined as species that virtually never cache food.
- Moderate cachers were those that cached food throughout the course of a year and cached many different types of food, but were never entirely
dependent on cached food sources for survival.
- specialized cachers were defined as species in which individuals cached a large number of items, the cached items were typically one food type, caching was seasonal, and cachers often recovered their items after long periods of time had passed.
Information about caching ability was then mapped onto the phylogeny of the group.
De Kort and Clayton’s findings that the ancestral state was moderate caching suggest that some corvid species evolved into specialized cachers, while others lost the caching trait altogether.
Learning and Foraging
Animals often learn how and on what to forage. Learning sometimes manifests itself in foraging innovations; brain size often correlates with the frequency of such foraging innovations.
The variance in foraging behavior seen across killer whale populations may be in part due to differences in
socially learned food preferences. (© Tory Kallman / Shutterstock).
Because the forebrain appears to be involved in behavioral plasticity, ethologists have hypothesized that larger forebrains might be associated with superior learning abilities (R. Byrne, 1993; CluttonBrock and Harvey, 1980; Dunbar, 1992; T. Johnstone, 1982; Jolicoeur et al., 1984; Mace et al., 1981; Wyles et al., 1983).
In North American and British Isle birds, and in the weighted mean of North American and British samples, the relative frequency of foraging innovation is positively correlated with the ratio of forebrain mass to brain stem mass (Lefebvre, 2011; Lefebvre et al., 1997a,b; Overington, Griffin, et al., 2011).
FORAGING INNOVATION AND DIVERSIFICATION IN
EMBERIZOIDEA
Taxonomically, Darwin’s finches are part of the avian family Thraupidae, which is embedded within the superfamily Emberizoidea
Darwin’s examined finches and the exquisite fit between the shape of the bill and the type of food eaten in these fifteen species (Lack, 1947). But Darwin’s finches, along with their relatives, have recently shed light on the phylogeny of flexible foraging strategies and rates of evolutionary diversification, defined as an increase in species composition over evolutionary time.
Social Learning and Foraging
Which new food items are safe, and which are
dangerous?
Louis Lefebvre and his colleagues Luc-Alain Giraldeau and Boris Palemeta ran a series of experiments to examine whether cultural transmission plays a role in pigeon foraging behavior.
Pigeons needed to learn to pierce the red half of a paper covering a box of seeds. The graph shows average latency to eating for four groups: NM (no model), BI (“blind” imitation), LE (local enhancement), and OL (observational learning). Pigeons in the NM and BI treatments never learned to feed in the experimental
apparatus. The quickest learning occurred in the OL treatment.
When a group member finally opens a tube with
food in it, the food spills on the floor and is accessible to all. Out of sixteen pigeons, only two learned to open tubes (these were the producers), while fourteen acted as scroungers.
