Witold Rybczynski's "green case for cities"

Writing in US monthly The Atlantic, architect Witold Rybczynski outlines his "green case for cities", arguing that there is too much focus on flashy green gadgets and not enough on practical building methods:

Putting solar panels on the roofs doesn’t change the essential fact that by any sensible measure, spread-out, low-rise buildings, with more foundations, walls, and roofs, have a larger carbon footprint than a high-rise office tower—even when the high-rise has no green features at all.

He also has a pop at the media for encouraging the "green gadget" trend:

Architectural journals and the Sunday supplements tout newfangled houses tricked out with rainwater-collection systems, solar arrays, and bamboo flooring. Yet any detached single-family house has more external walls and roof—and hence more heating loads in winter and cooling loads in summer—than a comparable attached townhouse, and each consumes more energy than an apartment in a multifamily building. Again, it doesn’t really matter how many green features are present. A reasonably well-built and well-insulated multifamily building is inherently more sustainable than a detached house. Similarly, an old building on an urban site, adapted and reused, is greener than any new building on a newly developed site.

Rybczynski leaves his most important point to the end, suggesting the future could be developments that are "dense without being vertical". He cites Montreal as an example of city where the dominant form of housing is a three or four-storey apartment block that doesn't require elevators.

Rybczynski makes a good point - while high rise is often touted as the solution to unsustainable urban sprawl, tall buildings typically require elevators, artificial ventilation (due to increases wind speeds and noise associated with openable windows at height) and heavy, high embodied energy structural components - perhaps three or four storey "walk ups" represent an ideal compromise.

But without solid figures, it's impossible to be sure. Does anyone know of any studies out there examining the energy footprint of different building forms?