What we see here is a pattern of drainage which is essentially dendritic, and which has its origins far back into the Tertiary and maybe even earlier. The key components of the uplands and lowlands have thus been in place for many millions of years, so that when the Pleistocene glaciations occurred the "grain" of the country, and the alignment of these old drainage routes, have strong influences on the patterns of both glacial action and fluvio-glacial channel cutting. Much meltwater was directed into these old valleys -- as we can see in the case of the Gwaun-Jordanston valley system, where subglacial channels have been deeply cut into the floors of older shallower valleys. The remnants of these old valleys can still be seen as "shoulders" high above valley floors. In some cases the positions of interfluves have been moved, and in other cases (as with Cwnm Gwaun) they have simply been lowered, so that we now have humped long profiles on channel floors. (This is a good diagnostic feature for a sub-glacial meltwater channel.....)
As I argued many years ago, and as now seems to be acceptd by the new generation of Geomorphologists, these channels are MUCH older than the Devensian. they have been used over and again, someimes by meltwater flowing subglacially and sometimes by water flowing subaerially -- in spillways from the glacial lakes of the area.
Is all of this relevant for our understanding of Stonehenge? Not really, but there are implications here for our understanding of landscape evolution at Rhosyfelin.