ANALYSIS OF MORPHOLOGICAL VARIABILITY IN NATURAL
POPULATIONS OF FISH
I. Introduction
In the laboratory sessions to follow, we will be
keying out fish and lumping many individuals into a single category called a
species. We will look on the species as a fairly homogenous unit. This, of
course, is not completely true. Variability exists between individuals of a
given species not only morphologically, but also physiologically and
behaviorally as well. The causes for such variability are both genetic and
environmental.
Taxonomists are interested in this variability because it tells them a great
deal about the population of individuals that comprise the species under study.
For example, if we plot the frequency of occurrence of some characteristic such
as number of anal rays, or number of scales in the lateral line, or length of
the base of the dorsal fin and study the resultant curve we can often say
something about the selection pressure operating on the population.
The length of the base of the dorsal fin has been plotted against the frequency
of occurrence in three isolated populations; A, B, and C. The base of the dorsal
fins in population A shows very little variation from the mean. Most individuals
have a dorsal fin whose length is about the same, suggesting strong selective
pressure against those individuals whose dorsal fin deviates very far from the
mean, a case of stabilizing selection.
Conversely, population C has quite a broad range with considerable deviation
from the mean, indicating that the selection pressure is not nearly as strong on
them as it was on population A. The curve for population B is skewed to the
right suggesting strong selection against those fish with a shorter than average
dorsal fin and relatively little selection pressure on the larger finned fish.
This is often called directional selection. Therefore, it should be clear that
the study of the variability of characteristics could shed light on some
fundamental aspects of the life of fishes.
Today we will look at the variability present in a natural fish population and
attempt to analyze it.
II. Procedure
1. From the jars at the front of the table, obtain
90 individual fish (10 at a time) from either population A or B. Sort into three
categories: male, female and juvenile (30 of each type).
2. For each individual measure its total length, the length of some body part
and count some repetitive unit
3. Record individually, keep juveniles, males and females separate.
4. Put data into a database (excel) to use for later population analyses.
4. Compute mean (x) and variance (S2) for both the meristic (counted
characteristics) and non-meristic (measured characteristic).
0 = S x S2 = S x2 – (S x)2 /N
N N-1
Where: x represents each individual measurement or count
N = number fish used
N S x S x2 0 s2
Male
Female
Juvenile
5. Run a t-test to determine whether the values for
the males differs significantly from that for the females and whether either of
these differ from the juveniles at the 95% level. What would you conclude?
t - test results:
Meristic Non-meristic
Male vs. Female
Male vs. Juvenile
Female vs. Juvenile
6. Prepare a scatter diagram for both the meristic
and non-meristic character plotting it against total length (must be done on a
computer). Now determine the regression line for each plot as well as the
regression statistic (r2). The curve represents a moving mean and the scatter of
points the variability. Are there any noticeable trends? Can you explain them?
Is there a relationship between the variables?
7. Compute the coefficient of kurtosis (Ks).
Kurtosis is a measure of the deviation of an observed frequency distribution
from a normal curve. If Ks is zero the distribution is the same as a normal
curve and it is said to be mesokurtic. If Ks is negative and distribution is
flatter than the normal curve and said to be platykurtic. If Ks is positive the
distribution is more peaked and said to be leptokurtic.
Compute Ks for all of your categories.
What would you conclude?
8. Define skewness and then compute skewness (Sk)
for all of your categories. When a distribution is symmetrical mean and median
are equal, thus Sk would be zero. Negative values indicate skewness to the left;
positive values indicate skewness to the right. Usually, a value of Sk which is
larger than +1 is indicative of significant skewness. What would you conclude
from your results?
9. Compare your results with that of a student using
fish from another population. Do you notice any differences: Can you explain
them? Write up your conclusions and turn in with your calculations at the next
lab period.