From an evolutionary perspective, an individual's success, or fitness, is measured by its relative contribution to the gene pool of the next generation (Begon et al. 1996, Krebs 2001). Because reproduction is the only way that an individual contributes its genotype to future generations, natural selection should favor organisms with genetically-determined traits that increase their likelihood of surviving to sexual maturity (Krohne 2001). These adaptive traits should therefore become and remain common in a population or species.

Predation is among the most common causes of pre-reproductive mortality in animals (Begon et al. 1996, Krebs 2001). Consequently, evolutionary forces have produced an impressive variety of predator-avoidance adaptations. Specific adaptations include speed that enables prey to outrun predators, or chemical or physical defenses that make an organism unpalatable or difficult to handle (Riessen 1992, Holzinger and Wink 1996, Lingle 2002). Organisms lacking these defenses must rely on different strategies to avoid being eaten, many of which are unusual and not immediately obvious. For example, snails in the family Physidae are slow-moving, edible, and have thin shells that make them quite vulnerable to predation by fishes and other shell-crushing predators (McCollum et al. 1998, Turner and Montgomery 2003). However, many physids possess a chemosensory system that is used in combination with behavioral strategies to reduce encounters with predators (Covich et al. 1994, Turner and Montgomery 2003). Specifically, many physids recognize the nearby occurrence of fish and other shell-crushing predators through chemical cues produced by crushed snails or the predators themselves, then move to habitats that provide refuges from predators (Stewart et al. 1999, Turner and Montgomery 2003).

Physa acuta (family Physidae), an ideal snail for this experiment.
Photograph by T.W. Stewart.

In this activity, you will gain an improved understanding of the complex ways that predators and prey interact in ecosystems, and how predator-avoidance behaviors help ensure coexistence of prey and predator populations and persistence of entire biological communities. You will also increase your familiarity with essential components of experimental design, and will use statistical analysis techniques to answer research questions objectively.

Through assigned readings (Covich et al. 1994, DeWitt et al. 1999, Stewart et al. 1999, McCarthy and Dickey 2002, Turner and Montgomery 2003), lectures, and class discussions, you will be provided basic biological and ecological background information pertaining to characteristics of common predators (fish) and prey (snails) of aquatic ecosystems. Additionally, your instructor will introduce basic elements of experimental design (hypotheses, replication, standardization, use of statistics to make objective conclusions), and show you available resources (prey species, predator species, stones, aquaria) that can be used in the experiment.

After this laboratory activity and relevant background information are introduced, you and some classmates will form a small group to design an experiment that can answer the questions:

  1. Can snails use chemical cues to detect threats to their survival in the form of shell-crushing predators?
  2. Do snails then move to habitats that provide refuges from these large predators?

After a pre-determined period of time, your instructor will ask each group to describe its proposed experiment to the entire class. Based on group responses, your instructor will devise a common experimental plan that the entire class will follow. All members of the class will be responsible for collecting, analyzing, and interpreting data that are required to answer research questions listed above.


Materials and Methods

Study Site

This experiment will be conducted in the laboratory or classroom. Snails used in the experiment were collected from the wild, or are captive-reared offspring of wild snails. Fish, if used, were captured from the wild or obtained from a fish hatchery.

Overview of Data Collection and Analysis Methods

Form a small group with other students. Together design an experiment to answer the questions:

  1. Can snails use chemical cues to detect threats to their survival in the form of shell-crushing predators?
  2. Do snails then move to habitats that provide refuges from these large predators?

Be prepared to describe components of your experiment's design to the entire class. Together, we will then design a single experiment that incorporates experimental design components of various student groups.

For example, one possible experimental design consists of two treatment levels, hereafter referred to as predator-cue and predator-free treatments. In the predator-cue treatment, you would record numbers of snails occupying aquarium habitats where they are vulnerable to fish before and after adding water containing chemical cues produced by fish or crushed snails. In the predator-free treatment, numbers of vulnerable snails would also be recorded before and after adding water. However, water added to aquaria of the predator-free treatment would not contain chemical cues produced by crushed snails or fish.

Below I describe data collection and analysis methods for the example experimental design described in the previous paragraph. You and your instructor might need to modify these methods to fit the experimental design used by your class.

    Setting Up the Experiment

    The following steps should be taken at least 24 hours before beginning the experiment. Therefore, your instructor might have completed these procedures for you.

  1. Place ten 19-L aquaria at evenly-spaced locations throughout the classroom. Aquaria should be distributed so that 2-3 students can observe activity in each aquarium. Fill each aquarium with dechlorinated water, but leave a 5-cm space between the water line and the top of the aquarium. If dechlorinated water is unavailable and tap water must be used, you will need to add chlorine-neutralizing solution (e.g., Aquasafe water conditioner®) to each aquarium because chlorine kills snails and other aquatic animals.
  2. Construct underwater refuges for snails by arranging patio stones, or similar objects, in a pile on each aquarium floor. The instructor might also provide ceramic tiles to be placed on top of the pile of stones. Each aquarium must contain the same volume of stones (e.g., 1 L), tiles (e.g., 2 tiles; 15 X 15 X 1 cm dimensions), or other objects.
  3. Carefully place 20 snails in each aquarium. Snails should have shell lengths between 3 and 6 mm. Shell length is measured from the spire tip (apex of coil) to the extreme tip of the aperture (shell opening). Snails in the 3-6 mm size range are large enough for you to see and small enough for many fish to eat.

    A snail (Physa acuta) from the family Physidae. Distinctive features of physids include a thin, coiled shell with a high spire, and an aperture (shell opening) that occurs on the left side of the shell.
    Photograph by R.T. Dillon, Jr.

  4. Fill two additional large aquaria (e.g., 38-L capacity) with dechlorinated water. One of these aquaria will be the source of chemical cues for aquaria constituting the "predator-cue" treatment of the experiment. Water in the second 38-L aquarium will lack these chemical signals, and this water will later be added to aquaria in the "predator-free" treatment of the experiment. Aerate water using a pump, airline tubing, and airstones.
  5. If fish will be used to generate chemical cues in the predator-cue treatment, release healthy fish into one of the 38-L aquaria. Fish placed in this aquarium should be individuals that have been previously observed to feed on snails. Your instructor will have already determined sizes and numbers of fish needed in this aquarium from preliminary trials that generated strong chemical cues. Add nothing to the second 38-L aquarium, or to either aquarium, if fish will not be used in this experiment.

    A pumpkinseed sunfish (Lepomis gibbosus). Pumpkinseed and redear sunfish (Lepomis microlophus) prey on physid snails and generate chemical cues that induce habitat shifts in these snails.
    Photograph by J.M. Haynes.

  6. The following steps should be taken 10-30 minutes before the experiment begins. Students should be present and participating in these procedures.

  7. If fish are not used in this experiment, simulate fish predation by crushing a large number of snails between your fingers, and placing them in one of the 38-L aquaria present in the classroom. If fish occur in one of the 38-L aquaria, feed them a large number of snails. Your instructor will predetermine the quantity of snails that must be crushed or fed to fish to generate a strong chemical signal. Again, add nothing to the second 38-L aquarium.
  8. Divide into 10 groups of approximately equal numbers of students. Each group will be assigned a number from 1-10, and will be responsible for managing one 19-L aquarium containing living snails and reporting data from it. The five odd-numbered student groups will manage aquaria of the predator-free treatment, and the five even-numbered groups are assigned to aquaria of the predator-cue treatment.
  9. Examine snails in your aquarium. Note snail behaviors, including crawling speed. If any dead snails are observed, remove and replace them with living snails of similar sizes.
  10. The Experiment

  11. At the instructor's signal, count initial numbers of snails in your aquarium that occupy habitats where they would be vulnerable to fish. Snails are considered "vulnerable" if you can see them on aquarium walls or floors, or upper surfaces of stones or tiles where visually-oriented fish could easily find and eat them. In contrast, snails inhabiting undersides of tiles, interstitial spaces between stones, or occurring above the water line on aquarium walls are considered "invulnerable" or inaccessible to fish. In other words, all non-visible snails, and those that have left the water, should be considered invulnerable. Report the initial number of vulnerable snails (X1) in your aquarium to the instructor, who will use data from all student groups to help you complete the second column of the predator-free and predator-cue treatment data tables located below.

    Students looking for snails at the end of the experiment.
    Photograph by T.W. Stewart.

  12. Complete Data Tables and Worksheets

    Follow steps 10-25 below and use these handouts:

    Predator-free Treatment Data Table and Worksheet
    [PDF] (45 KB) [DOC] (45 KB)

    Predator-cue Treatment Data Table and Worksheet
    [PDF] (44 KB) [DOC] (44 KB)

  13. After initial numbers of vulnerable snails are recorded, remove and discard 2 L of water from your aquarium. Remove water carefully to avoid damaging or disturbing snails. Students managing 19-L aquaria of the predator-cue treatment will now transfer 2 L of water from the 38-L aquarium containing fish or crushed snails to their aquarium. Students overseeing aquaria of the predator-free treatment will transfer 2 L of water from the 38-L aquarium without fish or crushed snails to their 19-L aquarium. Depending on results from preliminary trials conducted by your instructor, it might be necessary to repeat this procedure one or more times to produce a sufficient chemical cue concentration in the predator-cue treatment.
  14. After completing water transfers, observe behavioral responses of snails in your own aquarium, as well as aquaria under the care of other student groups. Do you recognize any behavioral differences among snails in predator-free and predator-cue treatments?
  15. Using scratch paper, record the number of vulnerable snails in your own aquarium at five-minute intervals. Report these numbers to your instructor. As snails in the predator-cue treatment begin to detect and respond to elevated predation risk, numbers of vulnerable snails in these aquaria should decline. When there are no further changes in numbers of vulnerable snails in aquaria of the predator-cue treatment, your instructor will give the signal to end the experiment. Report the final number of vulnerable snails (X2) to the instructor, and record class results in the third column of the data tables provided on the handouts.
  16. Data Analysis

    Now complete data tables and worksheets, and use paired-sample t tests (your class might use a different statistical test that is preferred by your instructor) to determine if initial and final numbers of vulnerable snails differed in predator-free and predator-cue treatments. Your instructor will assist with calculations and interpreting statistical analysis results. Additionally, if paired-sample t tests are used, you should also work through completed examples available in the Student Data section:

    Student Data Set #1: Previous class data and paired-sample t test results
    [DOC] (99 KB) [PDF] (72 KB)

    Procedures described below are adapted from Zar (1999). The box below provides additional information on paired t-tests.

      From: Zar, J.H. 1999. Biostatistical analysis. 4th edition. Prentice-Hall, Upper Saddle River, NJ.

      A paired-sample t test is used to determine the significance of the difference between two sets of paired data. In the experimental design described here, initial and final counts of vulnerable snails in each aquarium within one treatment are paired for the analysis. These pairings are based on our expectation that final counts of vulnerable snails in each aquarium should be affected by initial counts of vulnerable snails in addition to whether or not predator cues were introduced.

      Two paired-sample t tests are needed to test our experimental hypothesis that snails detect predator cues, then increase use of refuges from these predators. In the first t test we investigate for differences in initial and final numbers of vulnerable snails in the predator-free treatment. This first test is necessary to rule out physical disturbance associated with water transfer procedures as a cause for shifts in snail habitat use, and to separate this disturbance from chemical-cue effects in the predator-cue treatment. We should not find statistically significant changes in numbers of vulnerable snails in the predator-free treatment. However, in the second paired-sample t test (predator-cue treatment), we should find statistically significant declines in numbers of vulnerable snails following addition of water with chemical cues to aquaria.

    As the box above explains, we will need to perform two paired-sample t tests, one for the predator-free treatment and one for the predator-cue treatment. First, use the following procedures to complete the predator-free treatment data table and worksheet, and obtain t test results. We will then use the same procedures to obtain t test results for the predator-cue treatment.

  17. In the appropriate space on the predator-free treatment worksheet, enter the number of replicates (i.e., number of student groups or aquaria; n) constituting this treatment.

    Your answer should be "n = 5" if five student groups each managed a separate predator-free aquarium.

  18. Calculate difference values (d) for each pair of observations (d = X1 X2) in the treatment. For each aquarium or student group, subtract the final number of vulnerable snails (third column of data table; X2) from the initial number of vulnerable snails (same row in second column of data table; X1). Enter results in the fourth column of the data table.

    Sum the difference values (Σd), and enter the result in the final row of the fourth column of the data table.

  19. Calculate the mean, or average, difference value (d) from the difference values (d) present in the fourth column of the data table. Show your work and enter the result in the worksheet.

    d = Σd ÷ n

    Where Σ = sum

  20. Now subtract the mean difference value from each individual difference value.

    (d - d)

    Enter these values in the fifth column of the data table.

    The sum of these values, Σ(d - d), will be 0 if all calculations were done correctly.

  21. Square each value in the fifth column of the data table (d - d)2, and enter results in the sixth column.

    Sum values located in column six, Σ(d - d)2, and enter the result in the last row of the sixth column.

  22. Determine the number of degrees of freedom (DF) in this treatment. This value is equal to the number of replicates (i.e., student groups or aquaria) of this treatment, minus a value of 1. Enter results in the appropriate space on the worksheet.

    DF = n - 1

  23. Now use results from previous calculations to determine the variance of difference values (s2d).

    s2d = Σ(d - d)2 ÷ DF

  24. Now calculate sd, the standard deviation of difference values.

    sd = √(s2d)

  25. Calculate sd, the standard error of the mean difference value.

    sd = sd ÷ √n

  26. Finally, calculate the t statistic (t) for this treatment.

    t = d ÷ sd

  27. To determine if initial and final numbers of vulnerable snails differed, we must compare the t statistic calculated above to a critical t value. The appropriate critical t value for any test is based on 1) the significance level chosen (α), or accepted probability of erroneously concluding that paired samples differ when they actually do not (in biological investigations, this is usually α = 0.05), 2) the number of degrees of freedom in the data set, and 3) whether a one- or two-tailed test is to be used. Critical t values for one- and two-tailed tests and a 0.05 significance level are available in the table below. To identify the appropriate critical value for a paired-sample t test, first identify the row in the table below that corresponds to the number of degrees of freedom you previously calculated (DF = n - 1). Then read across the table to find the appropriate critical t value, dependent on whether you conducted a one- or two-tailed test. The decision to use a one- or two-tailed test should be made before conducting the experiment, and is based on student hypotheses of snail responses to predator cues. For example, a one-tailed test should be used if students predicted fewer vulnerable snails to occur in the predator-cue treatment at the conclusion of the experiment than at the beginning of the experiment (or vice versa). Alternatively, a two-tailed test should be used if students hypothesized that initial and final numbers of vulnerable snails would differ, but that snails could either increase or reduce their refuge use after exposure to predator cues (e.g., students might reason that refuge use could decline after exposure to chemical cues due to snail flight responses). If you have 4 degrees of freedom, critical t values are 2.13 and 2.78 for a one- and two-tailed test, respectively. Record your critical t value on the worksheet.

    Some critical values of the t distribution (from Zar 1999)
    Degrees of freedom (DF) Critical t value (α = 0.05; one-tailed) Critical t value (α = 0.05; two-tailed)
  28. The absolute value of the calculated t statistic must equal or exceed the critical t value for you to conclude that there is a difference in initial and final numbers of vulnerable snails. What do you conclude based on comparison of the t statistic and critical t value, and an examination of the raw data you recorded? Write your response on the worksheet.
  29. Repeat steps 13-24 to complete the data table and worksheet for the predator-cue treatment.

Questions for Further Thought and Discussion

  1. Describe snail behaviors in each treatment used in this experiment. Did you observe any changes in behavior after completing water transfers from the large aquarium with or without fish or crushed snails? Explain.
  2. What do you conclude from results of this experiment? For example, was the experimental hypothesis that snails use chemical cues to detect predators and subsequently increase use of refuges supported? How did you arrive at your answer?
  3. What questions emerged from this study? In other words, what additional questions would you address in a future experiment focusing on predator-prey interactions between snails, fish, or other organisms?
  4. How do predator-avoidance adaptations affect prey and predator population dynamics in natural ecosystems? Use information from lectures and reading materials to create and support your argument.
  5. Many species, including the snail you studied in this exercise, avoid predators by inhabiting structurally-complex habitat. Loss of structurally-complex habitat is a common consequence of human-mediated habitat destruction. Use results from this experiment and information from readings and lectures to predict effects of habitat destruction on 1) populations of prey that used the lost habitat as a refuge from predation, 2) populations of predators that rely on the prey for food, and 3) other species that depend on the prey or predator for food or some other resource.

References and Links


Tools for Assessment of Student Learning Outcomes (written for faculty)

  1. Biology majors (general ecology course)

    A student's grade on this exercise will be based on three components: 1) the instructor's assessment of student effort and level of participation in the experiment (10 points), 2) completeness of data tables and worksheets, and accuracy of calculations (20 points), and 3) a short (approximately five double-spaced pages) paper written in the format of a professional scientific journal (70 points). Modified versions of in-class discussion questions from this activity will also appear on an upcoming lecture or laboratory examination. See Questions for Further Thought and Discussion for examples of possible exam questions.

    Using data collected by a previous class, the instructor will demonstrate how to record data needed for statistical analysis. Your class will also work through two examples of paired-sample t tests so that you will become familiar and comfortable with calculations needed to generate t statistics. Your instructor will also help you interpret results of statistical analyses. Refer to the example:

    Student Data Set#1: Previous class data and paired-sample t test results
    [DOC] (99 KB) [PDF] (72 KB)

    and re-read the help box above for additional assistance.

    Class time will also be directed to scientific writing methods. The class will discuss each major section of a scientific paper, and what should be included in each section. A handout and checklist describing scientific manuscript components is provided in an appendix linked below to supplement lectures and discussions on this topic. In particular, you should periodically review the "Checklist for a Scientific Manuscript" as you write the paper. Your instructor will use this checklist to assist in grading your paper.

    Appendix 1: Handout on scientific writing and checklist of scientific manuscript components
    (adapted from Morgan and Carter 1999 and McMillan 2001)
    [DOC] (49 KB) [PDF] (75 KB)

  2. Non-science majors (general biology course)

    Non-science majors are evaluated on 1) an assessment of level of participation in the experiment (10 points), 2) completeness of data tables and worksheets, and accuracy of calculations and interpretations of results (20 points), and 3) the quality of written answers to questions based on this experiment (20 points). Your instructor will prepare you for answering these questions through in-class discussions of experiment objectives, results, and ecological applications of these results.