HOUSING FORM AND LAND USE : NEW RESEARCH
by Richard MacCormac
This essay first appeared in the RIBA Journal, November 1973
The advantages of houses are that they offer easy access, garages can be integrated with the dwelling, and private gardens can be provided for children, as can related public open spaces for the easy supervision of young children.
There are other equally important but less tangible advantages of houses. They continue a social tradition which is peculiarly English and classless, and local authority tenants’ houses need not be distinguishable from those of private owners – unlike local authority flats, which inherently symbolise the intervention of government in the housing process. Houses create a more friendly environment than flats through the opportunities for social contact offered by streets, and through private gardens and garages associated with the houses. They also offer a more easily maintained environment, because the house and it surroundings are identified with the people living there.
The net density of terrace housing can be surprisingly high. Fig 1 shows a possible arrangement of parallel three storey terraces of 3.66m frontage with 120 per cent car provision, giving a net density of 440 pph. Measured in this way, the density of the terraced houses in stage 1 at Tharnesmead, for example (fig 2), is more than 375 pph (150 ppa), which is half-again as much as the overall density of the town. Conventionally, it is not possible or even desirable to have such high densities, because of the lack of view and open space which such an arrangement would entail if it covered a whole site. But Lionel March’s mathematical demonstration of the advantages of court rather than street forms indicates one way to avoid this result (fig 3). He argues that, for a given angle of obstruction, the floor space index of an arrangement of courts is greater than that of an arrangement of streets, and therefore one can deduce that courts will achieve the same density as streets with larger spaces between buildings (fig 4).
When this theory of built form is implemented in actual housing design, a problem of vehicular access becomes apparent, because an arrangement of courts has to be broken into to admit vehicles, and therefore loses density. So a modification of the court form has to be devised which reconciles vehicular access with the maximum continuity of building. Figs 5 and 15 show such an arrangement, which originated in the London borough of Merton architect’s department. Here, the perimeter of the 16 acre site (6.5 hectares) is developed
as a series of alternating vehicular and pedestrian spaces with an unbroken terrace, which maximises length like an intestine.
It is interesting to compare this site and its three storey courts (fig 6) with another site in Camden – Burleigh Road (fig 7) – where the density is about the same (250 pph) but the buildings are largely medium rise. The Burleigh Road site could have used the perimeter principle (fig 8) and so limited the building height to three storeys.
If a site is very much deeper than these examples, you may ask how the court idea can be used without vehicles breaking through to reach the centre of the site. One answer is to insert another land use, which should be open space (school or park land), into the centre, as at Pollards Hill in the London borough of Merton (figs 9 and 10). Here, 33 acres of housing (13.3 hectares) have views and access to seven acres of park land. Vehicular access is limited to perimeter culs de sac. There is segregated pedestrian access to shops, and the density is sustained at 288 pph (115 ppa).
It is again interesting to compare this with stage 2 of Thamesmead (fig 11): a mixture of towers, middle and low rise housing, and two primary schools. The proportion of housing to school land is about the same as the proportion of housing to park at Pollards Hill (5:1), and thus a similar perimeter scheme would be possible there (fig 12), giving all families houses with gardens.
There is a further statistic to do with Thamesmead which maybe relevant to other local authority and new town developments. The combined sum of recreational and educational land in the original master plan of 500 acres (200 hectares) was almost equal to the land allocated to housing: 600 acres (240 hectares). If we take the initial perimeter diagram (fig 5) and string the 16 acres of housing around the area of adjacent school land, the layout unfolds like a concertina, and every house has a direct view and direct access to open space (fig 13). At least theoretically, it seems possible that Thamesmead could have been a garden city composed of 16 acre (6½ hectare) parks, containing pitches and playing fields, much like Parkers Piece in Cambridge. The contrast with the concentratedly urban image of Thamesmead is obvious (fig 14).
If housing adjoins recreational land, it may no longer be necessary to provide public amenity space within the housing area : either more houses can take up this space to give a higher density, or the density can remain the same and greatly enlarged private plots can be provided (fig 16). Because private gardens adjoin open space rather than other buildings, they can be built on, and the community is free to alter the boundary between public land and individual plots. This is a new and useful concept of ‘ribbon development’. The ideas discussed here originated in research at the centre for land use and built form studies in Cambridge, and have been developed in local authority work. ClearIy, research and practice can sustain each other more in this way. Here are some of the issues raised:
Though the cost yardstick increases for densities above 250 pph, it fails to compensate for the loss of amenity when dwellings are raised off the ground. If we agree that houses are more desirable than flats and can be built at densities up to about 250 pph, should not this be the top limit of densities, and why in the London area do architects continue to suffer tortuous design problems with densities of 340 pph (136 ppa)?
Conversely, if houses can be built at densities up to 250 pph, is there any good reason for continuing to build flats and medium rise buildings up to this density, as many authorities still do?
Both theory and practice show the advantages of considering several land uses simultaneously if the best housing is to be achieved. Planning should surely take housing layout strategies into account at its inception.
The separate administration of parks and school land in local government tends to perpetuate the very problems of high density housing which these recreational land uses were supposed to ameliorate. How can the organization of local government overcome this? Perhaps the concept of the community school is crucial here.
If good high density housing relies on the association of housing with recreational land, the suburbs might be the best place for such building.
It appears that ribbon development gives the largest plot size for a given density, but the diagram shown (fig 16) assumes a substandard access road with car access directly from it. Perhaps we need to challenge so called ‘highway standards’ if we are to gain the best overall advantage for housing itself.