ANIMAL BEHAVIOR LABORATORY

NOTES FOR TEACHING ASSISTANTS

SCHOOLING IN FISH

SUPPLIES

• 40 Zebra danios, 40 tiger barbs, and 40 serpae or other tetras
  

• five 7.5 gallon tanks filled and dechlorinated
  

• rulers to measure the size of the tanks
  

• stopwatches
  

• jars with lids
  

• wide, clear packing tape to attach jars to the tanks
  

• Fish Schooling Data Sheets

PROCEDURES

This experiment involves transferring fish to and from students' aquaria many times.   It is thus important to take some simple steps to minimize traumatizing the fish.

• To minimize handling (and hence traumatizing) fish during labs, separate the fish into their jars before lab(s) begin each day.   Have the students trade jars of fish rather than catch fish from the tanks.
  

• Subject fish (the freely swimming fish) should be used as subjects only once per day.
  

• Turn off the heaters in the aquaria holding fish for this lab a week before the lab.   As a result all tanks will come to the same (room) temperature.   Tiger barbs, in particular, are highly sensitive to changes of temperature.
  

• Keep acquaria and jars containing fish clean and filled with aged water (allow tap water to stand over night so chlorine will evaporate and temperature will equilibrate).
  

• Follow procedures for maintaining healthy fish .

POINTS TO EMPHASIZE IN THIS LAB

• releasers, releasing mechanisms, and sign stimuli
  

• simple statistics for analysis of a basic experiment

INTRODUCING SIMPLE STATISTICS

This lab provides an ideal opportunity to introduce basic statistic concepts and some simple statistical tests. Choice tests are best analyzed with Sign Tests or Wilcoxon Tests. Wilcoxon Tests are the more powerful, but also are somewhat more complex to understand.

Students should understand

• the null hypothesis
  

• the rationale for accepting a 5% possibility of error,
  

• observations with two possible outcomes (such as the result of tossing a coin),
  

• paired observations (such as the same subject responding to two different stimuli)
  

• the concept of a critical value in a statistical table

These concepts are explained in the statistical sections of the lab manual.

Note that you could analyze a coin-tossing experiment (or our fish-schooling experiment) with the Chi-squared Test. It is not the most natural test here, however.

The Sign Test is a much clearer test for observations with binomial distributions. The Wilcoxon Test is also suitable and is more powerful, because it takes into account the magnitudes as well as directions of differences. The Chi-squared Test is best introduced when you get to the foraging experiment with jellybeans.

All of these ideas are introduced in the statistical sections of the lab manual.