Laboratory Exercises in Animal Behavior - Squirrels and Food Selection
Squirrels and Food Selection
When we were in college the main pets on campus (legal ones that is) were fish and squirrels (Sciurus sp.). Obviously, the squirrels were everybody's pets because they roamed around campus while the fish roamed around a tank in someone's room. The squirrels became very active right before the weather got cold because they were collecting and storing food. Animal behaviorists refer to this behavior as "scatter-hoarding" because the animals collect scattered items and bury (hoard) them for future use. In this exercise, we are interested in identifying the "rules" by which squirrels scatter hoard. Choices animals make while foraging (trying to get food) are aspects of the "foraging strategies" individuals follow. Foraging strategies would seem to have obvious effects of survival and reproductive success, so we will assume, for now, that foraging strategies affect survival and reproductive success. Foraging strategies are also interesting to animal behaviorists because choices made by individuals often have both positive and negative consequences, which must be balanced against one another within a particular strategy.
In this exercise, you will investigate two aspects of squirrel scatter hoarding. First, do squirrels prefer to forage where food is plentiful or where food is less plentiful? And second, do squirrels prefer to forage near cover or far from cover? You will also investigate the interaction between these two aspects. If faced with a choice, do squirrels perfer to forage where food is plenitful and cover is far away, or where food is less plentiful and cover is nearby? In other words, which is more important, the amount of food available or the distance from cover?
We will use observations combined with controlled comparisons in an attempt to study the scatter-hoarding behavior of squrrels. The class will be divided into research groups, and each research group will be responsible for one hour of observation. When all observations are completed, you will combine (pool) your data.
Study design.-Observations will be conducted at four patches of ground, each 10 x 10 meters in area. Two of the patches will be on open ground 100 meters from trees or other suitable squirrel cover and the other two will be 10 meters from cover. (The distances are somewhat arbitrary-you can use other distances as long as one is clearly close and the other far from cover.) The four patches should be as similar as possible in all respects except for distance to cover. The patches will be arranged in two sets, each set containing a patch near cover and one far from cover. In one set (choose which set by some random procedure such as a coin flip) the close patch will get 200 shelled peanuts (peanuts with the shells removed) at the beginning of each one-hour observation period and the other will get 100. The opposite arrangement will be used for the other set. The patches may be within sight of each other, but should not be close together. The reason for four patches is so we can see the effects of amount of food (200 versus 100 peanuts) and how these effects interact with distance from cover (100 versus 10 meters).
Data aquisition and analysis.-We will assume that patches close to cover are safer places than patches far from cover, and that patches with more peanuts are better than patches with fewer peanuts. Our data will be the number of peanuts eaten and storred in each patch during each one-hour observation. Each group will gather data on all four patches during a specified hour of the day. At the start of each observation the four patches must be stocked with the appropriate number of peanuts. At the end of the hour, the number left in each patch must be counted and the numbers recorded. These numbers should allow you to determine whether squirrels prefer areas with lots of food (high density food areas) or areas with low food density, and whether they prefer to hunt for food (forage) near cover or far from cover. Pay close attention to the time of day when the most food is removed in each patch.
Presentation of results.-One of the most important things that scientists do is to present their results to other scientists. They do this for two reasons: (1) to inform other scientists of their findings, and (2) to allow other scientists to evaluate the design of the study, to agree or disagree with their interpretations of the results, and to replicate (repeat) the study if they choose to do so. One way that results are reported is on a poster at a session of posters. You will present your results on a poster. Keep in mind that a poster is a visual presentation of (1) the topic you studied, (2) why you studied it, (3) the methods you used, (4) the results you obtained, and (5) your interpretation of those results. Keep in mind also that because posters are visual, it is important for them to be visually attractive-a good poster will draw your attention from across the room, and will present its information quickly and clearly.
Things to Think about
Topic.-We are studying the foraging strategies of squirrels. What kinds of "strategies" might squirrels follow, and how might each help a squirrel to live longer and/or reproduce more? Each one of these possible strategies is a functional hypothesis. What hypotheses are you testing, and how can each be supported or disproven? What effect might number of peanuts and distance to cover have on a squirrel's survival or rate of reproduction? What would happen if two aspects of a strategy or two different strategies conflict with one another (for example, if the patch with the best food was also the patch with the highest risk of predation)? We made some assumptions about safety and food quality. How might these assumptions be tested? Do you have any data to test the assumptions? How might you determine whether a particular foraging strategy is really functional (whether it really improves survival or reproduction)?
Methods.-Why was it necessary to flip a coin to decide which patches would get lots of food? What steps do the groups need to take to make sure that each group does the same thing, and why is that important? Would it be best to keep the same arrangement of patches, or to choose a new arrangement at the start of each hour-long observation? Is it sufficient just to count the number of peanuts taken during the hour, or should groups also observe the behavior of individual squirrels? If you decide to observe squirrels, what will you observe, how will you record it, and how can you make sure that all groups do the same things?
Analysis.-Once you have your pooled data, how can you analyze them? Remember, we are trying to identify the foraging strategies that squirrels follow. What possible strategies could squirrels follow? How do your data either suppor to refute the possible strategies? What functions might the possible strategies serve?
- meter tape or other measuring device
- chalk or string to mark the areas.
- 1,000+ peanuts
- poster board, graph paper, and markers
Tips for Teachers
Remember that there are no "right" and "wrong" hypotheses (although some will be more reasonable than others in that some will make sense given what we know about squirrels and foraging behavior, and some will not). What is most important is for your students to realize that any possible strategy can be tested as a scientific hypothesis, according to the "scientific method." The hypothesis should be based on existing information (e.g., what we already know about squirrels and their foraging behavior), and it should explain something about our topic of interest.
For example, we know that squirrels are scatter hoarders and we suspect that they hoard food to improve their survival during the winter. We might therefore hypothesize that squrrels adopt a strategy to gather and hoard the most peanuts in the shortest time. This "peanut maximizer" hypothesis would predict that squirrels always go to the patch with the most peanuts regardless of the distance from cover (i.e., regardless of the danger).
An alternative hypothesis might be that the risk of predation is more important than foraging efficiency, so that squirrels would adopt a strategy to keep the risks as low as possible. This "predation minimizer" hypothesis would predict that squirrels always go the the patch closer to cover regardless of the number of peanuts it contains (i.e., regardless of the food density). Note that each hypothesis makes a clear and testable prediction, which will be either supported or refuted by the results of our study. Note also that it is possible to get results that partially support both hypotheses. We might find, for example, that squirrels try to find the best balance between maximizing peanuts and minimizing predation.
One of the most difficult lessons for students to learn is that critical hypothesis testing should attempt to disprove ("falsify") hypotheses. For example, the predation minimizer hypothesis could be disproven if squirrels always avoided the patches close to cover (as long as our assumption about safety is correct). A simple game can serve to illustrate the point. After you introduce this topic and explain about hypothesis testing, tell them you are going to have them play a guessing game to learn some methods of hypothesis testing. Tell them their job is to guess the "rules" you use to choose numbers that you'll write on the board. Then write three numbers on the board (e.g., 1, 2, 4). Tell them the game works as follows. A student can ask you whether a particular number fits your rule. You will answer truthfully, either "yes" or "no." If the answer is "yes" then you will add that number to the list on the board. If the answer is "no" then you will put that number in a separate place on the board. The questions you answer are analogous to the observations and experimental results generated by research, and the two list of numbers are the "data" the students have. Once a student thinks s/he knows the rule, s/he will "publish" it by describing the proposed answer to the class. The class will then discuss and evaluate the proposed rule. Anyone who disagrees with the "published" rule may publish one of her/his own by immediately proposing it to the class. Throughout these discussions, you will not say whether a proposed rule is correct, nor will you volunteer the true rule. These discussions are analogous to the process by which scientists communicate their ideas with others, and by which competing hypotheses are evaluated. If a proposed rule wins ready acceptance by the entire class, the proposer could win the Nobel Prize, but if the proposed rule is rejected by the class or disproved by a new "experiment" (by proposing a new number), the proposer's reputation is damaged and s/he must "sit out" for some short period (e.g. five minutes).
Note that the numbers in our example seem to follow an obvous rule: double the last number. In reality, the rule is: "positive integers" (1, 2, 3, 4, etc., but not zero and no fractions). In almost every case, students will pick an obvious rule and then propose numbers to you that fit the obvious rule. So, you are likely to have students propose 8, 16, 32, 64, etc. Let them keep going for a while; you keep saying "yes" and adding the next number to the list. Now, let's think about what the students are doing. In one respect, they are testing hypotheses. They have a particular rule (hypothesis) in mind, and they are proposing numbers (making predictions) that can be tested (you say "yes" or "no"). Unfortunately, they will never get any closer to identifying the rule because they are attempting to "prove" their hypothesis instead of trying to "disproving" it. What's the difference? All of their numbers FIT their hypothesis. You will say "yes" and add the next number to infinity (or at least until the bell rings).
In order to test the hypothesis critically, a student must propose a number that DOES NOT FIT the hypothesis. Why does that help? Because if the proposed number does, in fact, fit the rule, then the student has learned that the hypothesis is INCORRECT; the student has disproved the hypothesis. In contrast, even if your students double the last number one hundred times, they still haven't proven the "doubling" rule because other rules can still explain the numbers in the list. It's only by eliminating some possible rules (hypotheses) that we make any progress. So, in our example, if the student has in mind "double the last number" then s/he should propose a number that doesn't fit the proposed rule. For example, after 64, the student could propose 63. What happens when you then say, "yes, 63 fits!" Obviously, the rule is NOT "double the last number." In contrast, if the student proposes 128, you would also say "yes" but there are still other rules that also fit (e.g., a number larger than the last).
Interestingly, the process of hypothesis testing in animal behavior is exactly the same as in this number game (well, almost the same). If we contine to propose "tests" that simply confirm what we already know, we don't really get anywhere. It's only when we propose a test that does not conform to what we already know that we make any headway-we learn what explanations DO NOT APPLY to our behavior of interest.
Your students may come to the realization that testing to confirm hypotheses isn't getting them anywhere. If they do, be sure to point out their discovery and give them lots of reinforcement to this truly significant realization. If, however, they are still confirming hypotheses after several rules have been proposed, you can take a "time out" to discuss the differences in strategies of hypothesis testing with your students, then let them try out their new "method of disproof." (Note that this is the method advocated by noted philosopher of science, Karl Popper). Let them then resume guessing and you should begin to list lots of numbers that DO NOT fit your rule. Once they begin to get the idea, stop the game and resist the temptation (and the frequent requests) to tell them the rule. Remind them that this is science-we can NEVER find out the rule by asking a "higher authority."
Students should be encouraged to spend time thinking carefully about the design of the study and what potential results might mean. We have suggested a "balanced, unreplicated" design in which the two factors, number of peanuts and distance from cover, are pitted against each other with each combination (200-close, 200-far, 100-close, 100-far) represented once only. Other designs are possible, and each will have advantages and disadvantages. Make sure that each group discusses possible designs, and that all groups agree on a single design (because they will pool their results). Ask them how they will ensure that all groups use the same methods, and if they choose to do observations of squirrel foraging, how they will know that all groups use the same observational methods and criteria. Remind them that they will have to pecify all of their methods and criteria when they present their results.
Although your students will be gathering some very simple data (counts and perhaps some observations), these data will enable your students to answer some good questions. Point out to them that science is not fancy, expensive equipment, but a process of investigation. Make sure your students know ahead of time how they will use the peanut counts to test for effects of food density and distance from cover. How does the "balanced, unreplicated" design enable them to determine which factor, or combination of factors, affects a squirrel's decision of where to forage (when and how long to forage, and what to eat are other questions that are of interest to animal behaviorists). Have them think of all of the possible results (e.g., no differences, more taken from 200 regardless of distance, more taken from close regardless of number, etc.) and how each possible result either supports or refutes their hypohteses. Clear thinking about what results might mean before the results are in hand is a very important component of hypothesis testing and the scientific method.
Once the groups have gathered their data, leave the discussion of results to the groups themselves. You can monitor their discussions just to make sure they aren't completely off base, but give them lots of rope at this point. You should discuss poster presentations with them, however. Have each group prepare its own poster, and don't let the groups "compare notes" until their posters are finished. Tell your students that "pictures" are more effective than "words" in posters, so if they have a choice between a graph or a table of results, they should use the graph; if they can show the design with a picture rather than a written description, they should use the picture. Tell them also that a good poster grabs your attention and then holds it. Too much written material will tend to lose readers, but not enough will be uninformative.
Finally, have your students participate in a "poster session" in which all the posters are displayed, and each group in turn describes its results. Even though all groups will be using the same data, you will find remarkable diversity in their presentations of the results, and especially in their interpretations of results. Have the groups critique each others' posters, emphasizing that critiques should be constructive but thorough-what is good and not so good about each poster. What works and doesn't work in each poster, and how can each be improved? If your poster session works well, you should consider "advertising" its success. Invite other students, and even your principal and members your School Board to view the posters and talk about the project with your research groups. In the short term, you'll bring credit to your students and your course. In the long term, you might improve your funding for supplies and equipment, and you might allay some of the board's fears about teaching evolution.