The key factors distinguishing the southeast (where the Anglo-Saxon kingdoms formed) from the West and North (both of which remained more culturally and linguistically 'British') are the topography and agricultural practices. The southeast is Britain's lowlands, and it was densely settled farmland used to grow cereal crops (wheat, etc) when Britain was part of the Roman empire. The west and north are uplands, and they were sparsely settled, with less good farmland, less grain production, and more investment in livestock.

Here's a map showing high and low elevation, by Chris Greene from Oxford. Wales isn't included on this map, but it would be mostly purple (high elevation, and less land under the plough).

Here's a map of 'Anglo-Saxon' graves (graves with weapons in them), made by a grad student at the Univ. Florida.

And here's the two maps overlaid (same source). Note where the 'Anglo-Saxon' graves are: in those lowland valleys, the good land for growing grain. See where they (mostly) aren't? The uplands. There are none at all in the mountains of Wales.

Here's another one, with the dots replaced by a shaded heat map showing areas with the greatest density of material (the red areas). This maps shows even more clearly that 'Anglo-Saxon' graves are most common in those fertile, lowland valleys.

Why did these two cultures (one 'Germanic', one 'Celtic') emerge where they did? Anglo-Saxon England and British Wales (and the Northwest) fall into two naturally different ways of inhabiting and working the land, rooted in topographical and agricultural differences in their physical landscape. These differences existed in Roman Britain long before it was divided into the early medieval kingdoms, and continue to shape the emergence of political and cultural differences in these regions through the present.

The real question, then, is not why these two regions split apart along the lines that they did - the line itself is structured by geography. The real question is, why did the new identity, the Angles and the Saxons, emerge in the good land, and why did the old identity - Roman and British - end up in the hills?

Our historical sources provide the answer we're very familiar with, that it was a violent conquest. Bede (writing in the 8th century), the Anglo-Saxon Chronicle (9th century), and an earlier sermon by a pastor named Gildas (the only contemporary source we have, from maybe the mid 6th century) all say that the Saxons stole the best land and pushed the British west. But archaeology gives us a much more complicated, and perhaps more interesting, picture.

There's archaeological evidence that people did migrate to England from Germany. Culturally, there are new cultural practices emerging in the easternmost parts of Britain in the early fifth century, at sites like Spong Hill, where the local well-to-do start burying their dead in a massive cremation cemetery. Their burial practices look a lot like those from across the Channel in Germany, and nothing like the burial practices being practiced by the native people of Roman Britain. Recent genetic studies have variously estimated anywhere between 10-50% of England's population descending from German immigrants, so there were a lot of people coming over (though a recent study suggests this migration lasted for centuries, and wasn't a one-off invasion like the old conquest story would have us believe).

But these immigrants weren't the only people making changes in the southeast in the fifth century. In other parts of lowland Britain, Roman life continues, but it restructures: people move out of towns, farm less land (but raise more variety of crops), stop importing pottery, and start pulling down old villas, recycling them for building materials, and living in new wooden houses. Many of these people are locals - important Roman men who change with the times and make new opportunities for themselves by taking over local resources and setting themselves up as strongmen. These strongmen also emerge in upland Britain - across the island, politics are becoming more divided and more local. James Gerrard, The Ruin of Roman Britain (2013) is a great overview of these political changes; Harrington and Welch (2014) are the best recent discussion of the economic changes.

How did these local strongmen and newcomer immigrants interact? The evidence points overwhelmingly toward peaceful mixing. It's not especially surprising - immigration was common in the late Roman empire, and people were used to moving around. Recent studies have found a number of Roman Britons who were born in North Africa, testifying to the cosmopolitan experience of living in a Roman colony.

There are several ways to study these interactions. The first is the total lack of evidence for violence: during the period of supposed conquest, we find no evidence for Roman villas being burned down, little evidence of violence in people's graves - indeed, the burials with weapons that we associate with the conquering Anglo-Saxons mostly don't appear until two generations after they supposedly started conquering and enslaving the locals. The huge 'Anglian' cremation cemetery I just mentioned, Spong Hill, had hardly any weapons in it from the fifth century.

But the more compelling evidence is the mixing of people and culture in the cemeteries. You can tell where someone grew up by studying the stable isotopes in their teeth, and this is showing that people buried with culturally 'Anglo-Saxon' artefacts were often born in Britain, while people who clearly immigrated often looked like locals. In fact, in a new cemetery (RAF Lakenheath) that looks like it should be packed full of late fifth-century Anglo-Saxon warriors, located right next to Spong Hill (the huge 'Germanic' cremation cemetery), everyone in the cemetery seems to be local - no immigrant Angles or Saxons at all, despite the fact that the cemetery was in use beginning right about 450 when the Angles and Saxons were supposed to have landed, and despite the fact that all the stuff in people's graves was culturally 'Anglo-Saxon' (Hines, forthcoming). Genetics is painting a similarly complicated picture, in which people of continental ancestry lived, ate, married, and died alongside the Romano-British populations.

What does this mean? It means that things in England were complicated, varied a lot from site to site, but in almost all cases involved immigrants mixing in with the local population.

So: how do we get from there, to the highlands having a Romano-British culture, and the lowlands an Anglo-Saxon one? And why do our historical sources say it was violent if the archaeology shows peaceful mixing / intermarriage?

The answer comes down, ultimately, to a question of myth making and, again, geography.

In south England and Kent, the easiest trading partner was Gaul. In East Anglia and the northeast, it was easier to trade with Saxony and Scandinavia. In the west, however, it was easier to go down toward Spain and trade Byzantine and Mediterranean goods. These regions of trade created communities and networks, with elites like those buried in Sutton Hoo (in East Anglia) trading with elites in Sweden, while those in the west, like Tintagel in Cornwall, imported Byzantine pottery (full of wines and oils). (Cf. Williamson 2013 for geography and ethnicity; Harris 2003 for trade between Byzantium and W. Britain)

The strongmen who tried to rule these kingdoms also interacted with each other, often struggling for each other's land and the loyalty of each others' people, and they needed to define clear boundaries and differences between each other. How, precisely, these histories and identities formed is difficult to trace with precision (because so few written sources survive before 600), but we know that by the seventh century kingdoms most closely connected (geographically) to Germany had begun to emphasize the German part of their ancestry, while kingdoms still politically connected, albeit in the loosest possible way, to Byzantium defined themselves as the heirs of Roman Britain. These myths formed the basis of moralizing sermons (like stop-sinning-or-God-will-not-save-you-from-your-Saxon-neighbors preacher Gildas), and for origin stories like those recorded in the Anglo-Saxon Chronicle, or the contemporary and equally politicized chronicle of the Welsh historian Nennius. These stories rewrote the largely peaceful processes of local cultural blending in the fifth century to fit the power struggles of the new political dynasties that emerged in the sixth century, and explained England's geographic and cultural differences as the products of battles won by the famous ancestors of these kings who needed to make names for themselves.

Firstly, the notion of a "conquest" of Britain by the Anglo-Saxons has largely been dispelled. The expansion of Anglo-Saxon style burials in southeast England is now thought to be indicative of cultural assimilation by Britons into the new Germanic society. To be a wealh, or foreigner/slave/Briton, was undoubtedly a lower place in Anglo-Saxon society. So it is possible that Briton families, wishing to escape these social conditions, might have practiced Anglo-Saxon style burials with continental dress and grave goods.

STBs or "sunken-feature-buildings" were some of the main new building styles of the fifth and sixth century, along with timber-style rectangular buildings, usually with at least one separate room. Differing from the continental longhouses and Romano-British villas, these STBs are most likely indicative of further hybridization between Saxon and Briton culture. Even further, there is evidence that many Anglo-Saxon settlements were constructed on the sites on former British towns, in contrast to the antiquated assumption that they avoided former Roman settlements.

So why did this cultural assimilation/conquest not take place in the west? Chiefly, the Anglo-Saxon kingdoms in the 500s were small, disorganized, and generally unable to carry out state functions. By 700, we see the decline of burial goods and rise of high-class settlements to show us that the Anglo-Saxon kingdoms were becoming more centralized.

By contrast, some early Celtic kingdoms were quite centralized. We have evidence of the kingdom of Rheged (modern Lancaster area) being able to rally large numbers of warriors from Pictland and Wales as late as just after 600 AD. Strathclyde in southwestern Scotland was able to resist the Anglo Saxon kingdom of Bernicia and further enroachment until 1034, and was notable enough to be featured in Irish chronicles around the 700s. Gildas makes reference to Dyfed and Gwynedd as being ancient kingdoms with sustained dynasties. Gwynedd was strong enough in the 600s to raid the English Midlands and Scotland regularly. There was also likely more centralization among British kingdoms in Wales, because it is difficult for historians to determine when they ceased being Roman provinces and began being kingdoms. In 600, the British kingdom of Dumnonia likely stretched across Cornwall, Devon and Somerset; at almost 200km long this would dwarf most Anglo-Saxon kingdoms of the time. Dumnonia also had an unbroken line of single kings, which made them unique among British kingdoms. There is even some linguistic evidence that Dumnonian kings ruled on both sides of the English channel, including Celtic Brittany to the south. To be fair, by 700 this kingdom had shrunk to just Cornwall, but it is indicative of the size and centralization of the British kingdoms the Anglo-Saxons were fighting against.

So it is likely a combination of Anglo-Saxon political disorganization and the relative political centralization of the British kingdoms that prevented the Saxons from culturally assimilating or conquering the island in its entirety. We see as the 700s continue that increasingly centralized kingdoms like Wessex, Mercia and Northumbria emerge, and these were able to push further and establish the boundaries that we today understand between England and Wales and England and Cornwall. (Also, these centralized kings do not include the Irish and Picts. They were likely ruled at the time by loosely organized "overkings" who would command local chiefs)

Source: Cambridge Medieval History volume 1, chapters 9 and 10