Re: Why do dominant alleles 'take over' recessive alleles?

Actually, saying that dominant alleles "take over" recessive alleles is
misleading.  A dominant allele shows up in the phenotype (the observable
properties of the organism), and in such a way that when an individual is
heterozygous (that is, has one dominant allele and one recessive allele),
the phenotype is the same as you'd see in an individual that has two
dominant alleles.  The important thing to remember, though, is that the
recessive allele is still present in the heterozygote, and it hasn't been
changed; it can be passed on to the next generation, and any offspring with
two recessive alleles will still show the recessive phenotype.  

Why is a dominant allele dominant?  That is, when there are two alleles
present, why does the dominant allele seem to control the phenotype?  

Let's consider what genes are, and what they do.  A gene is a piece of DNA
that codes for a protein.  That is, it can be transcribed, resulting in a
messenger RNA, and that mRNA can be translated to make a protein.  There
are several types of proteins, but the most important groups would be
structural proteins, such as keratin that is found in hair and fingernails,
regulatory proteins that control expression of other genes, and enzymes. 
Of these, the largest group is enzymes.  Enzymes are protein catalysts;
almost all chemical reactions that occur in cells are catalyzed by enzymes.  

Often, a dominant allele codes for a functional enzyme.  Many recessive
alleles code for non-functional proteins.  Even a change of one amino acid
in a protein can cause it to fold wrong, so that its active site is
altered, for instance.  This can result in an enzyme with a different
function, but probably the most likely outcome is a protein that just
doesn't do anything.  

Consider this example:  One allele of a gene in rabbits, call it "A", codes
for an enzyme that makes the amino acid tyrosine into the brown pigment
melanin.  Another allele, "a", codes for an inactive protein.  An AA
homozygote makes lots of the enzyme, produces melanin, and has brown fur. 
An aa homozygote doesn't make any enzyme, so it produces no melanin, and
has white fur.  Now consider the Aa heterozygote.

The Aa heterozygote has some active enzyme, and is able to perform the
chemical reaction and produce melanin as a result.  Maybe it doesn't have
as much of the enzyme as you'd find in a AA homozygote, but it still gets
the reaction done.  Remember, an enzyme is a catalyst, so it can perform
the reaction over and over, so even with less of the enzyme, in many cases
it's possible to convert all the available reactants (tyrosine, in this
case) into products (melanin).  The result is that the Aa individual makes
melanin and has brown fur, just like the AA homozygote.  We'd say that A is
dominant. 

This isn't the only possible way for two alleles to act in a heterozygote,
of course.  Sometimes the heterozygote doesn't produce as much product as
the homozygote (so maybe you get a tan rabbit), or sometimes two alleles
might code for enzymes that catalyze two different reactions, and the
heterozygote could have two different products.  These are examples of
other forms of dominance.  

To sum up: genes (and their specific forms, alleles) code for proteins.  A
heterozygote has two different alleles of a gene, coding for two different
proteins.  The heterozygote's phenotype depends on the chemical effects of
the two different proteins it has.  Often, one allele codes for an active
enzyme, and as long as you have some of that active enzyme, you get the
same phenotype; in these cases we say that allele is dominant.