Ensatina Model Guideline
Natural selection can favor body coloration that conveys information to a potential predator if it protects an individual from being eaten. Previously, we sampled the coloration of natural populations of Ensatina and Taricha from a number of regions (representing different sub/species). We hypothesized that differences we saw in coloration between Ensatina species reflected two strategies for avoiding predation. Many of you then proposed a possible experiment you could run in order to test this hypothesis. The Ensatina model experiment is just that.
For this short type of report (1-2 pages text + figures or tables if necessary) do not use the traditional section headings included in longer reports. You should briefly introduce the question and logic of what you did to show your understanding of the material. Focus most of your attention to presenting your results. In this short type of report, you should present your results and discussion/interpretation together. When doing this, take care to present the material in a logical manner so that the reader does not get lost in the statistical details of your results, but focuses on the interpretation of those results and the connection between each of your results. This is in fact how Nature and Science papers are written, so we are not leading you astray in your writing techniques!!
Analysis:
The data from model transects are measures of the number of models of each type that died:
|
Body |
|
Eye |
|
|
|
Black |
Yellow |
Total |
||
|
Brown |
16 |
7 |
|
|
|
Orange |
4 |
8 |
|
|
|
|
Total |
|
|
|
You should analyze this data using a Chi-squared test comparing the number of individuals that survived against the number that survived. There were 107 models of each type, so the number survived is just 107-#died. You should determine if the number of attacks on each model type is random with respect to color type or if some colors provide more or less protection. To do this, you must calculate a chi-squared expected probability shown below for each of your comparisons (only one is done for you below):
Expected value: (row X column) / (grand total)
So for Brown-Black:
| Body Color |
Died
|
Survived
|
Total
|
| Brown |
16
|
91
|
107
|
| Orange |
4
|
103
|
107
|
| Total |
20
|
194
|
214
|
So for your expected value of Brown that died you would get: (20 X 107)/214 = 10 (as you would for Orange that died).
You should then do the same for the number that Survived
You then calculate the Chi-squared statistic using the formula:
| S |
(Observed – Expected)2 |
|
|
(Expected) |
You then compare this number to the Critical Chi-squared
values to determine your p-value (look up the values in the Coot Foraging
Chi-squared calculator). The interesting comparisons are primarily brown-black
against the other types (3 pairwise tests BB vs BY, OB, BY) -- why?? You should
also compare which is more important: eye color or body color. To do this,
pool the results for hits on eye color ignoring body color and the same for
body color ignoring eye color (ex: for eye color you should have two values
- one for black the other for yellow). You should then run the chi-squared
on these values in the same manner as model type comparisons.