This diagram (from the modern bird molecular clock paper [Free PDF!] I quoted earlier) is our answer – click to enlarge it, so you can see the labels clearly. The most important thing to notice is that T(morphology) always comes later than T(gene). That means that the origin of a new morphological feature has to come after the origin of the genetic toolkit that is involved in its development. The genetic divergence is invariably earlier than the morphological divergence.
And this is exactly what we see when we compare the fossil record (T(morphology)) and the molecular clock analyses (T(gene)). And from that, you can guess how I answer to the question of whether molecular clocks and fossils are compatible: yes.
What we have are two different data sets. They are undeniably linked together, but that doesn’t mean we can mix them as we please. We don’t know the genetic make-up of any extinct organism. All that data comes from recent organisms; we compare their genomes and see which gene sequences are conserved. We find out what those genes do and assume that since they’re conserved, they must have been present in the last common ancestor.
We know that morphology comes from genetics; what we don’t know is how a change in genetics will affect morphology. Those conserved genes may have had completely different functions when they first originated, and later got co-opted to their modern functions – this is the biggest danger one can run into when basing a molecular clock study on developmental genes and it’s another reason why the molecular clock cannot give us any divergence times for species.
Basically, what we are measuring is not the origin of species, it’s the origin of genes. And simply having a gene does not make a species.
(Note that I’m avoiding going on a tangent about what ‘species’ means and whether it’s a useful concept!)
That genetic difference reflects evolutionary distance is undisputable. That genetic difference is positively correlated with time is also indisputable. But claiming that the rate of genetic change is constant in time is bullshit.
The variations in rate are, in fact, impossible to predict (some might even say impossible to approximate!), and the current methods of the molecular clock are based on too many false assumptions to actually be accurate – we are not measuring any exact dates. And those dates that we do get are not species divergence dates, but genetic origination dates.
Molecular systematists need to refine their methods.
Both sides need to understand what the molecular data tells us.
Palaeontologists need to understand developmental biology and genetics, just as molecular systematists need to learn about the fossil record.