The Iron Age is not simply a time of roundhouses in various forms. Rectangular structures of four or more posts are often found on cropmark sites, and interpreted as granaries on no strong evidence; one could equally construct a roundhouse from a four-post structure. The nature of these structures remains a major concern e.g. Dunwell 2007, 61-62).
There is a variation in circularity, with a number of structures notably oval, while some structures in the Atlantic zone have been seen as D-shaped 'semi-brochs', a concept supported strongly by MacKie (e.g. 1991; 2008, 267, 274-5), but rejected by others, who seek to explain these as eroded and collapsed roundhouses (Harding 1984).
The Atlantic Late Iron Age appears to show increasingly diverse range of structural forms. These are predominantly cellular forms, but there are also a small but significant proportion of rectilinear structures like the wags (stalled buildings) and similar structures (Baines 1999, Cowley 1999), some Argyll duns , and buildings external to earlier brochs, e.g. at Dun Vulan (Parker Pearson et al. 1999; Gilmour 2002, 2005). Questions over the function of many of these buildings are yet to be satisfactorily resolved. There is also some continuing use of circular forms (e.g. late wheelhouses at Scatness, Dockrill et al. 2010 and forthcoming). But how significant is the difference? Should wags be seen as aisled roundhouses transformed into rectangular form (Harding 2009 , 276)?
Beyond the Atlantic zone, the evidence of buildings in the first millennium AD is minimal after c. AD 200, and has even been suggested to represent 'tableaux of desertion' (Hill 1982b, 10). There is some evidence of rectilinear forms emerging (Pitcarmick type houses, Anglian halls). Cellular forms also occur (e.g. Ardestie; Harding 2004, 240-2), and the continuation of round and oval forms is in evidence (e.g.Buiston, Crone 2000; Easter Kinnear, Driscoll 1997 and the circular 'homesteads' of Perthshire (Taylor 1990, although their dating may be earlier than he argued, cf Hingley et al. 1997). This reduction in evidence may also reflect a change to non-earthfast building techniques. Whatever the cause of this apparently sharp diminution of structural settlement evidence, it means that obtaining a coherent idea of the range of settlement forms and the social structure that lay behind them, is going to be profoundly challenging.
Recent work in East Lothian has suggested that here the development of non-roundhouse buildings was underway in the 2nd-1st century BC at the site of Phantassie (Lelong 2008b). This site also serves as a useful reminder of survey bias, as it was unrecorded prior to invasive fieldwork, and the cellular structures had no earthfast foundations. At Phantassie they survived because they used stone - but similar buildings of turf would leave no trace, and such 'invisible' architecture poses a serious challenge (Loveday 2006). For those who want a more hierarchical Iron Age, the landless peasantry may have lived in exactly such hypothetical turf or timber houses which would be a struggle to recognise today.
What forces led to the move away from roundhouse architecture in different parts of the country? The contexts, chronology and significance of the introduction of rectilinear forms of architecture in various parts of Scotland during the first millennium AD require a major input of future research and synthesis.
What roles were played by buildings such as wags, souterrains, four-posters, and the irregular-shaped buildings found outside some brochs?