Hey Shadowfiend,
Those threads are awesome! That's exactly the kind of flexibility I'm after. And very inspiring interviews too. I'd seen images of his "Neckling" work, but hadn't taken the time to read through all of it. Thanks for the heads up 
Ok, so I'm going through this quite slowly, trying to anticipate any potential problems in advance. I'm asking myself a series of question, so I'll post them here as I go along.
The first question is: Will I be able to use the topology of my current base head to make my new base head? Which leads me to ...
STEP 1: Collecting reference.
This is really important. I'm starting with a basic male head. From this I intend to make a female head, an old male head and a fat male or female head. I also want to create a fantasy head, and have the ability to blend between all of these head sets, much in the way that Taron does (see Shadowfiend's links). This means that they will all need to have the same topology, which will need to be based on the topology of the most complex head.
My second question is: Is my current base head ready for this pipeline?
STEP 2: Preparing the base head
My first step in this process will be to analyse the current base mesh and check for any aspects that could complicate the rendering later, like n-sided polys or unmerged edges. My aim is to have an all-quad head, but I don't have a problem with triangles as long as they're not in areas where a lot of deformation will occur.
This may seem like a waste of time if you're using a head set that's already been through a production, but don't assume that a head that has rendered smoothly in Maya Software will work just as well in Mental Ray with subdiv approximation displacement. I've made that mistake 
I'll use Polygons > Cleanup with the "Select and Cleanup Geometry" selected for "faces with more than four sides" to copnvert any n-sided faces into triangles.
At the same time I'll make any topological changes in Maya, according to the requirements anticipated from my reference, such as adding edgeloops, or extending the mesh. This I'll do while the base mesh is connected to all of the blend targets. Then I'll bake topology to targets to propogate this change (using the Deform > Edit Blenshapes > Bake Topology to Targets in Maya 7.0, but for those not using 7.0, there is a thread on www.jonhandhisdog.com that explains how to do this in earlier versions).
I should say, though, that I'm not making any change to the form, only structural changes to the mesh topology. The changes to the overall shape will be made in Zbrush and then added to the base head as a parallel blendshape.
Next is to layout the UVs well. This is crucial for bringing displacement maps back from Zbrush (but, if you're also texturing in Zbrush, I don't see any reason why the UV layout cannot be done there using, say, GUV tiles). I've laid out my UVs in Maya because I liike to be able to take the file into Photoshop and see what is going on. I still find it quicker and easier to create the colour and bump maps in Photoshop, and then touch them up in zBrush, but that's just me.
At this point I'll just briefly run the mesh through the pipeline to make sure that I haven't missed any problems. I'll Export it to zBrush, divide to max level, export .obj and displacement map back to Maya, quickly hook it up using the Zatt script (http://206.145.80.239/zbc/showthread.php?t=20802) and do a test render. Mental Ray will tell me if it is having any issues with the mesh.
Ok, enough for now. Next post will have some images to demonstrate these concepts.
Cheers,
-sp
...and Scott, you know we want to see you workflow 
