Mutations and creationist misunderstandings
Let's consider the types of mutations that can occur. Mutations are important because, as we teach our introductory biology students, mutation is the "ultimate source of variation." Natural selection, genetic drift, and gene flow can change the amount of variation in a given population, but can't produce variation de novo.
The simplest mutations are the point mutations. These mutations change a single base in a gene. As you recall, there are 4 kinds of bases in DNA, which are abbreviated A, T, G and C. DNA has two strands, and every A in one strand is matched by a T in the other, every G matches a C opposite it, and vice versa. Point mutations change one half of the pair, and then the natural duplication of DNA perpetuates the error.
Some point mutations rely on quantum phenomena, so can be said to be entirely stochastic (random). Electro-magnetic radiation hits and either causes free radicals (don't worry about it) to evolve from water, or directly damages the DNA strand. Free radicals can change a G-C pair to T-A. Free radicals are results of other common biochemical processes, as well, but I like the random changes from EM radiation.
There are also isomers (variant arrangements of atoms within a molecule) of bases which react oddly with polymerases (the enzymes which build a new DNA strand using the original strand as a template). These too produce single-base-pair mutations. There is also a natural decay of C (which pairs with G) into uracil (which pairs to A, usually it only occurs in RNA), resulting in a C-G to T-A transition. Other natural chemical decay processes can cause the polymerase to mix things up.
There are various repair mechanisms in the cell, but sometimes the repair removes the correct base and inserts the wrong one. That leaves the mutation in place on both strands.
Some mutations of a single pair are more complex. If the polymerase gets mixed up, it may skip a base. Since DNA is read in groups of three, removing one base causes the resulting protein chain to change radically and unpredictably. The same goes for an accidental insertion. These mutations are called frame-shift mutations, because the normal three base "reading frame" gets shifted one base over. Frame shift mutations don't have to be a single base.
There are more elaborate sorts of mutations as well.
In sexually reproducing organisms, a process called "crossing over" occurs. A pair of chromosomes line up, the DNA is broken and the ends of the chromosomes switch. Sometimes that happens right at the same place on each chromosome, but it can also be off, perhaps by quite a bit. That means one chromosome is now larger; the other is smaller. One chromosome may well have two copies of one or more genes, or a regulatory region may be changed or removed, causing changes in the expression of existing genes.
Viruses insert genes, and sometimes a mutation or an evolved anti-viral response disables some part of the viral gene. But the virus has added new material to the genome, possibly breaking existing genes or changing the structure and functional properties of the proteins they produce. Asexual organisms can also exchange genes by various means, with similar results.
In sexually reproducing organisms, the effect of genetic recombination can be a drastic change in protein structure from parent to offspring. The chromosomes may be unmutated, but interactions between the mother's genes and the father's genes can produce unexpected protein changes.
These sorts of large-scale rearrangements of the genome are well-documented, are a common component of introductory biology classes and certainly genetics courses. I don't doubt that Dr. Davidson knows that.
I expect that Paul Nelson misunderstood something that Dr. Davidson told him, but either way, the claim as presented is nonsense. In the text he quotes from Davidson's article in Science, Davidson at least mentions gene duplications, which is a good first step. I'll get deeper into that article in a later post. I just wanted to clear up what sorts of mutations are out there. Creationist ring-leader Michael Behe pubished a whole article about how evolution couldn't generate some sort of evolutionary novelty, but his whole simulation was based on single-base mutations, ignoring the rich diversity of mutation types, and rendering the exercise meaningless as far as actual biology is concerned. Perhaps he should start reading TfK.
Update: Blogger has been having issues, so I'm republishing this.