Let's start with some basic principles. In general, those lots which are easiest and cheapest to build on would be expected to be the most valuable. Approximately rectangular lots are easier and cheaper to build on than lots in irregular shapes, and within those constraints approximately square lots are easier and cheaper to build on than narrow and elongated ones.* We should therefore expect to see the residential areas of most cities composed of roughly square lots, and in fact this is exactly what can be witnessed in, for instance Japanese cities, with lots in the Tokyo development shown below of 32' x 38' (all images from Bing Maps):
To maintain a favorable ratio of private to public land, the roads in such a scenario must be quite narrow, as I mentioned in the previous post, and that is what is generally seen in Japanese cities.
In American cities, by contrast, the common historical pattern has been both wide blocks and streets. In order to respond to high demand for housing, therefore, the solution has been to slice lots thinner and thinner. Although this does increase the number of lots per block, it adversely affects their utility, leaving the builder construct a very long and narrow house. An example is from an older neighborhood in Buffalo, NY -- a neighborhood more than a century older than the one above -- with lots of 30' x 175':
Why, it could be asked, were these blocks made so wide? What sort of development was anticipated? Why were the blocks not subdivided with additional streets, as famously happened in Philadelphia? Instead of any of these options, it appears most lots were densified vertically, through the construction of two and three-family wooden houses.
And yet, even were each lot built out as a three-story, three-unit home, the density would be less than in the Japanese example, and without one single-family detached house. Not all blocks in Buffalo are so wide, but few are much narrower.
For comparison, let's examine the blocks of Detroit, a city long known for its prevalence of single-family detached housing. Lots here (a neighborhood of the 1910s or 1920s) are 34' x 125', a more reasonable dimension that's helped the city maintain its high share of single-family housing. Note also the presence of alleys here through the center of the block.
As the development of American suburbs progressed during the 20th century, the width-to-depth ratio continued to moderate. Here, in Levittown, from circa 1950, lots are 60' x 110' with no alleys. The blocks are visibly narrower than in the Detroit example:
Finally, in some of the developments of the 1960s and 70s we see what are approximately square residential lots. Here, for example, is Herndon, VA, with lots of 90' x 105'. Interestingly, the ranch houses still "sprawl" across the wide lot, continuing to create the visual effect of a more or less solid wall of houses to a person going past:
Having the distance between the side walls of houses as a fraction of that between the fronts and backs of houses is a common feature of American residential developments in all eras, it seems.
There appears to have been a reversion to Levittown dimensions in the subdivisions of the 1990s and 2000s, with the New Urbanism even reintroducing deep narrow lots in the pre-1950 style in its more recent developments.
I'd posed the question above of why the earliest lots were made to be so inefficiently long and narrow. Land values obviously compelled the narrow widths, but why had blocks been made so wide in the first place? Were the surveyors of the 19th century under the impression that American homeowners, like the villagers of Tsarist Russia or medieval western Europe, would be tending to "dacha"-style backyard gardens for sustenance?
Perhaps. John Reps, in his The Making of Urban America, confirms that some of the earliest planned settlements in North America intended individual house lots to be used for gardening purposes:
"If it can be tentatively concluded that the New Haven [Connecticut] plan, with its generous provision for open space, was no sudden inspiration of the moment, there is no ready explanation for the source of its form or dimensions. ... The large residential blocks did not long retain their original form. Intended to provide generous garden plots adjacent to residences, these deep squares were eventually divided into four smaller blocks by new streets running at right angles from the midpoints of their sides. . . ." The Making of Urban America: A History of City Planning in the United States, p. 130.Were the blocks of a city like Buffalo, laid out as early as 1804 by a man from rural Bucks County, Pennsylvania, similarly designed to provide an urban facsimile of the rural farmstead in what was then wilderness of Western New York? Whatever the reason, the blocks were poorly-suited to the industrializing city that emerged a few decades later. The legacy of these blocks with their very narrow lots was only gradually discarded over the next century or more.
Rather than building very narrow multifamily structures on these narrow lots, which in general have not fared well in Buffalo and which have simply vanished from large swathes of the city (as have, to be fair, many of Detroit's single-family houses), what other options are available? There are a few related possibilities, assuming we limit ourselves to single-family structures. For example:
- As in New Haven and Philadelphia, build new streets into the blocks to create new, smaller blocks with more reasonable dimensions.
- Allow condominium development to stretch back from the street in rowhouse and/or detached form, which would allow numerous owner-occupied units on each lot.
- Where alleys exist, allow new houses (i.e. ADUs, accessory dwellings, "granny flats," etc.) to be built along them.
Where three lots once hosted perhaps three duplexes with two two bedroom units each (the Buffalo "2/2"), there are now eight fairly large (1800 sq. ft.) single-family homes, or, if homes were divided by floor, as many as sixteen one and two-bedroom units. A small lane (or really narrow street if you prefer) runs in between the homes, and cars can be parked alongside houses. The backyard is heavily truncated, but there is enough space for a small patio. If the above design were extended to another three lots on the opposite side of the block, the lane would become a small through-street, helping to open up the large blocks. Further, if the houses were made slightly smaller, I think it would be possible to fit in ten rather than eight. You can do the math.
This style is largely what Nathan Lewis has called "single-family detached in the traditional style." As an infill strategy, it is uncommon in the older cities of the Eastern US, but may have promise in circumstances like those in Buffalo where deep lots leave a large amount of unusable land. A condominium development would look much the same, similar to what can be seen here in Stamford, CT on similarly deep lots:
Houston, in particular, has countless examples of this sort of townhouse or compact single-family infill development, especially in and around the Montrose neighborhood. A Google maps tour of the area will reveal the various ways such developments can be planned and arranged.
The ADU approach is so well-known I need not provide any examples of it here, but it is obviously less practical in situations where there are no alleys, as in Buffalo, which if lots were to be legally separated would require either complex shared driveway arrangements or the use of flag lots, impractical where the lot is already so narrow.
There is much more that could be done with multifamily housing, but that will need to wait for another post.
*As Nathan Lewis has noted, long, elongated houses are inefficient users of resources, requiring more wall length per square foot than a square house. These dimensions are also less favorable for energy conservation, as there are more points for heat to escape.