Actually it is easier to create morphs in a low res figure, less points to move around means smoother results particularly in the high res, and if you start with the high res there is no way to convert to the low res, Unfortunately even if the high res figure is a sub-d of the low res figure which I assume it is, that still does not mean that you will end up with a successful morph by simply converting the sub-d because each modeler has a different way of converting point order so you would likely need the precise modeler that the originals were converted in. than the next problem is, was the figure split? if it was was it done before or after the high res figure was created I would assume after otherwise their would be gaps in the mesh, that causes more possible vertex order changes particularly if it was poser that did the splitting.