The Ethogram: quantifying behavior and testing hypotheses

Basic questions about animal behavior come initially from observations. Observation leads to questions, which lead to hypotheses, which lead to predictions, which lead to experimentation. You must understand your study animal before you can begin testing ideas. At some point, someone actually must watch animals in their natural habitat. You will do this as part of your experience in this course

Ethology is the comparative study of animal behavior. Ethologists study the biological roots and meanings of animal actions. The first step in that process is to construct an ethogram. In its simplest form, an ethogram is a quantitative description of an animal’s normal behavior. Constructing a useful ethogram demands time spent watching animals, taking careful notes, and making sense of the observed behaviors. You end up with an annotated catalogue of behavioral patterns grouped in a coherent fashion that describe what a given species does in a given environment.

For this class, you will build an ethogram for an animal of your choice as part of the semester-long project. You will understand your animal to the extent that you can generate questions and testable hypotheses. You will then use the skills you learned while constructing your ethogram to test one hypothesis about your study animal.

Overview of responsibilities


Schedule of assignments


Choosing a study species

You should weigh several factors before choosing the subject species for your ethogram. Variation in behavior among individuals is the rule, can be expected, and is important. You will want the opportunity to observe several or more animals of the same species. So long as your presence does not change the species, behavior, you should try to observe at the closest range possible (and always downwind from mammals), in order to pick up the smallest details of movement. For some animals, you will need binoculars. For others, you will not.

You should choose an animal that can be reliably found and observed. It won’t make sense to study a Cooper’s Hawk (Accipiter cooperii) since they are rarely seen and only fleetingly when they are. You must also consider time of year. Ruby-throated Hummingbirds (Archilochus colubris) are fascinating and can easily be lured to a feeder. Unfortunately, hummingbirds have left central Ohio by late August.

Humans have had profound impacts on the behavior of animals. Ideally, you will study wild animals in their natural habitat. That isn’t always possible. For this class, the following hierarchy describes what I’d like to see:

    1. “Wild” animals in their natural habitat. This might be soldier beetles () in the fields around the environmental center, or pigeons in downtown Mount Vernon.
    2. "Wild" animals in large enclosures. What is “large”?. A 100-liter aquarium may be appropriate for an adult crayfish (Cambarus sp.), but not for a channel catfish (Ictalurus Punctatus).
    3. Free-ranging domestic animals such as dogs or chickens (Gallus gallus), or in large enclosures (e.g., horses, cattle, llamas). Highly intelligent pets (e.g., dogs, cats, etc.) of your own should not be used, for your presence will certainly affect its behavior.
    4. Domestic animals in small enclosures. For many domesticated animals, small enclosures are their natural habitat.
    5. "Wild" animals in small enclosures. The Columbus Zoo is a full of these. The problem with zoos is that they often engender aberrant behavior due to the nature of captivity. However, the Columbus Zoo has some exhibits that minimize this effect. The gorilla exhibit, for example, is not large with respect to the animal’s natural habitat, but may be of sufficient size and variety to provide you with suitable subjects.

You need to decide on a study species as soon as possible. The longer you wait, the fewer your choices for many animals won't be available after mid-October. You should consult with me before making a final decision on which animal to study.


How many individuals should I watch?

The best way to characterize an animal's natural behavior is to watch as many different individuals as possible, then summarize their “average” behavior. If you only watch one individual, you will know a great deal about that individual’s behavior, but very little you can generalize. You must, therefore, choose a species for which you can observe a number of individuals.

Ideally, you can tell individuals apart. That is not always possible. For example, you may study Canada geese (Branta canadensis) that live in this area. Without capturing and marking them, you will not be able to know whether you are always watching the same individual. One way around that it to watch a large enough group that the probability of watching an individual twice is low.

In general, I’d like you to choose an animal in an environment where you can watch at least 4 different individuals. That is a minimum number. You are better off watching 1000 different individuals for 1 minute each than a single individual for 1000 minutes.

Back to Overview

 

Catalog of behaviors:

In your field work, you first must characterize the range of behaviors you see. This includes describing all the different kinds of behavior you observe and the context in which they occur. Only after you have built a catalog of possible behaviors for your animal, will you be ready to quantify those behaviors in terms of frequency, duration, and context. In general, you need to be aware of the following goals when building a catalog of behaviors:


The catalog: names and definitions

Your catalog is a list of behaviors and explicit definitions for each. The catalog allows you to organize your data sheets to quantify behavior. A catalog lets you define with one word any particular behavioral pattern you are likely to observe when the time comes to build your ethogram. List all the behaviors you have observed and define each of them. For example, some grey squirrel (Sciurus
carolinensis
) behaviors might be...

Behavior Definition
resting No movement, eyes closed
freeze No movement, eyes open, erect posture
saltate Jumping locomotion, as compared to running
chirp Vocalization, short duration, quiet between pulses
chatter Vocalization, long duration and continuous


Once you have built the basic catalog with objective definitions, you can organize your catalog according to behavior types. In the above list, “resting”, “freeze”, and “saltate” might fall under a category called “movement”, whereas “chirp” and “chatter” might fall under “vocalization”. Such a hierarchy will let you organize your datasheets and your analyses more easily.

Back to Overview

 

Equipment and note-taking

An accurate, comprehensive ethogram depends upon keen observation and accurate, efficient notation. Although tape recorders and video cameras are great for capturing observations quickly, they are often unreliable and they demand that you double the time you spend taking data, since transcribing the data from the tape often takes as long as recording it in the first place. The simplest way to take data is to record notes by hand directly into a notebook with sewn-in or glued-in pages. Loose-leaf notebooks tend to lose pages in a strong wind. You should use a ballpoint pen with waterproof ink and a clear plastic shaft (e.g. BIC pens) so you can monitor the ink supply. If you are working in cold weather, pens may not work well without some way to warm them. A pencil is also adequate, but remove the eraser. Erasing field notes is a very poor idea because you may later wish to know what you had erased. A single line through incorrect notes conveys the same meaning and leaves all notes legible.

Alternatively, you can use a pen-based computer and a software called Ethoscribe (Tima Scientific, Inc.). The biology department owns a few of these systems and you can check them out from me this fall. They are not as simple as pen and paper, but they are a very efficient way to collect data if you are comfortable with computers.

Stop watches and binoculars can be borrowed from the department, as needed.


Field Notes (from T.C. Grubb, Jr.)

At the start of each observation session, you should note the subjects' physical environment (e.g., wind velocity and direction, temperature, percent cloud cover). Sketch a map of the study area emphasizing components influential to your work. Throughout each session, keep track of the time of day. Glance at your watch periodically and note the time in the lefthand margin of your notebook.

Behavioral data should be recorded truthfully and without preconceptions of what an animal should be doing. In modern ethological studies, behavior must be quantified. Phrases like “several times a minute”or “a few meters”, have no place in field notes. Using the metric system, always put number estimates on behavior; the whitetail deer walked 7 m in 5 sec, or the pigeon raised its wing 12 cm.

Field notes (and your thinking while recording them) should be couched in operational terms. Simply record what you see without ascribing motive or function to the behavior. This is an important and sometimes quite a subtle skill, but one you should cultivate.

Incorporate in your notes frequent drawings of behavioral patterns; stick figures are perfectly adequate. Besides giving you a pictorial record, drawing forces your attention to all aspects of behavior. Only in drawing the posture of a drinking grey squirrel might you realize that you do not know the position of the tail. Those with cameras can use this valuable tool to aid analysis and, later, illustration of the ethogram.

Back to Overview

 


Quantyifying behavior - Major Approaches: Focal vs Survey

Once you have completed your catalog of behaviors, you are ready to quantify how your animal spends its time. You are expected to observe your study animal for at least six hours. You should aim for at least four individuals (see above) to be observed in this time period. You should devote an approximately equal amount of time to each individual you observe. You should be careful about when you observe them. It is best if you can observe each individual multiple times (i.e, three twenty minute observations on the same individual is better than a single one-hour observation on that individual). You should distribute your observation periods through the day so that you can get a generally representative sample of behavioral patterns. Alternatively, you should restrict your observations to particular time(s) of the day and distribute your observations evenly within those constraints. You should use one of the following methods:

Focal approach. With this method, you locate a single individual and follow its behavior for a standard time (or as long as possible up to that time). If a focal individual moves out of your view, then you start a new sequence of observations on a new focal individual. Selecting the focal animal can be systematic (you might just follow young animals) or randomized (select a random number from a table, then follow the nth individual encountered). During a focal study, you should record the following data as they occur:

How you record your data will vary with the stage of your study. When building your catalog, you will have taken extensive notes. That won’t work when trying to quantify how an animal spends its time. Organized data sheets are critical for focal studies, because they allow you to record data quickly and efficiently during the observation period, and to tabulate the results accurately afterward. A well written catalog of behaviors will allow to create a form for collecting data for the quantification phase of the ethogram. An example data sheet is available online. We will discuss how to use this sheet in class. Please feel free to modify this (for example, you may prefer to give the first time of observation, then use a stopwatch to record the duration of each behavior).

Survey approach. This method involves watching many individuals at the same time. At systematic (or randomized) times, you count the number of individuals performing each possible behavior. For example, if you are watching a flock of ten geese in a lake, you might set your timer to alert you at one minute intervals. When the timer goes off, you note what each of the geese in the flock are doing at that instant, as well as the context and consequences of their actions. You will repeat this every minute for your entire observation period. Taking survey data in the field also is facilitated by creating a data form that can be filled in quickly. Ethoscribe software also allows you to use the survey method.

In consultation with me, you will decide which approach is appropriate for your study animal.

Back to Overview


Summarizing behavior: analysis

Two methods are commonly applied to express patterns in behaviors observed through these techniques.

Time Budgets (or energy budgets) can be derived from focal studies. A time budget indicates the amount or proportion of time that animals spend in different behaviors, or in performing different classes of behavior. To construct a time budget, you must add the time spent on each behavior for each individual squirrel. Then calculate the proportion of time spent on particular behaviors, out of the total observation time. When you are able to observe several different individuals, then each individual can be a "replicate" or independent estimate of time allocation to the behavior of interest. You summarize your data by doing a time budget for each individual, then calculating mean values for each behavioral measure of interest.


The second pattern that can be derived is called a transition matrix. A transition matrix expresses the probability that one behavior is followed by another. For example, we might expect that digging a hole is followed by burying an acorn with a probability of 0.85 (85% of the time), by chasing another squirrel with a probability of 0.15. A transition behavior links "chains" of behaviors and is a technique to look at relationships between different behaviors. In a transition matrix, behaviors are listed in rows and columns in a table. The "cells" of the table is the relative frequency that the reference (row) behavior is followed by the next (column) behavior. A simple exarnple of a transition matrix follows:

Action / Next Action
dig carry object bury object freeze chase
dig
0.15 0.8 0 0.02 0.03
carry object
0.4 0 0.55 0.05 0
burry object
etc.        
freeze
         
chase
         
           


Note that the rows must add to 1.00, as each behavior must be followed by some other behavior (assuming that al1 possible behaviors are included in the transition matnx).

By examining transitions over different individuals or in different conditions, you may learn about "complexes" of coordinated behaviors. Also, you can ask questions about how different conditions influence the flow of behaviors through time.

 

Here is a link to an excellent discussion of how to construct Time Budgets and Transition Matrices by Dr. H. Jane Brockman (U. of Florida)

Back to Overview


Hypothesis formation and tests

Once you have built a catalog of behavior, done a focal (or survey) study in which you have characterized your animal’s behavior in detail, and sumarized the data for general trends, you should have thought of a few questions and hypotheses. For example, if the grey squirrel were your study animal, you might have noticed that there are plenty of Red-tailed hawks (Buteo jamaicensis) in Gambier and that the squirrels are affected by their presence. You might hypothesize that squirrels must spend more of their time allocated to predator defense. Several specific behaviors might be interpreted as predator defense (note that this is now interpretation, not direct observation). For example, "freeze" or "run to tree" (when not in the context of being chased by another squirrel) might be interpreted as predator defense behaviors.

First, you might do a preliminary analysis of your ethogram data to see how much of the time budget is allocated to predator defense. The next step might be to do a more directed ethogram with which you could test this hypothesis. You might wish to compare the proportion of time allocated to these behaviors for squirrels in high- and low-predator environments. You already know how to do a time budget. All you need to do is identify areas of high- and low-risk of predation and compare squirrel time budgets between them.

This is the final phase of your project. In collaboration with me, you will formulate your hypothesis, plan your study, and build a final ethogram that tests your hypothesis.You will write a one-page document that presents your hypothesis and the specific prediction you will test, as well as what you will measure in this directed ethogram.

Back to Overview

 

Sharing your findings


Class presentation - the Poster Session

Ethogram presentations begin after break. The presentations will be presented in Poster format. Your poster should give the following information:

Obviously, with limited space, you don't have much room for any of these. Concentrate on the important points. Pictures really do tell a thousand words, especially pictures that are data driven. Use charts to communicate your results (the general time budget and the hypothesis test results).

You will need to load your PPT presentation on the P:Drive to have them printed. See the Poster Prep Instructions for details.

Schedule

To be determined after break.

 

Final write-up

You will write a final paper that presents your Ethogram project as a scientific paper. I have prepared a guide for Writing Scientific Papers that you should consult. It reenforces what you already know from reading the primary literature throughout the semester. The format of the paper should follow the Guide for Authors for the journal Animal Behavior (follow the link to Animal Behaviour, then choose "Guide for Authors"), with the exception that you should use American English ("behavior"), rather than British English ("behaviour"). When in doubt, consult one of the papers you have read for class this semester from Animal Behavior.

Your final paper is the culmination of the work you have done this semester. The general organization of your paper follows the above guidelines (Title Page, Abstract, Introduction, Methods, Results, Discussion, etc.). Within these general guidelines, you should keep in mind the following with regard to how you should integrate the two major phases of your project:

Background literature and length requirements:

The Biology 109 Resource page has good information on how to use Excel and Minitab to test hypotheses and present your data, as well as general guidelines for scientific writing.

 

Back to top of page