Via Jarret Walker I found this poster on the variety of American city grids.
It’s pretty cool, but I don’t think there’s as much variety as the poster claims.
See, beyond a certain point, you can have a really large “official” grid but it will be broken up into smaller pieces. Alleys will become commercial streets in their own right. Likewise, you can design out a teensy-tiny grid, but certain streets will become basically abandoned, no commerce, just back entrances to buildings and whatnot.
For instance, he cites Macon, GA, with block faces exceeding 400 feet. And it’s true that is technically the block size. But the alleys in Macon are commercial strips in their own right with a variety of retail and entertainment establishments that can’t be accessed from the front. In fact this “alley” isn’t too different in scale from a “street” in Virginia or Pennsylvania. Salt Lake is nominally even bigger than Macon, but the same “cutting up” reduces the effective grid size to something smaller. In fact, where Macon’s alleys cut its blocks in half, Salt Lake’s blocks are sometimes cut up into thirds.
In Philadelphia, this phenomenon has actually gone in both directions. William Penn’s original 17th-century plat laid out a city with a broad grid, intended to allow small-scale farming. As the place urbanized and industrialized in the horse-and-buggy era, this was cut up and cut up again until Philly was chock full of tiny blocks as little as 100 feet on a side. With electric traction, the city was able to vastly expand its borders, and the need for these little tiny streets died out. Drury Street is still hopping, but many others are just a place to store trash dumpsters, with walls knocked out so that buildings which originally fronted on the alleys now open onto the next street over.
Meanwhile, Houston Center converted a couple blocks of Caroline Street into what’s basically an alley, complete with loading dock. In the Syd Mead future this was supposed to be the fate of most of our cities. While it hasn’t really turned out that way, I’d wager that has less to do with Syd being wrong about the need for multi-block buildings, and more to do with the fact that most cities have turned out to be pushovers when it comes to closing public streets for mega-developments.
So, the way I see it, there are two truths:
(i) There’s a rough lower and upper bound to what size a block should be, outside of which streets will be cut up or abandoned, and
(ii) These boundaries are largely dependent on the size and transport modes of the city.
Now, for a guy who took vector dynamics AND fixed-income securities analysis, the natural next step is to create a mathematical model for this. Grab a bunch of data points and do some multiple regressions and voila! Except, I don’t think I could ever do that. Because, how do you explain New York?
Lower Manhattan has the small grid, it’s got blocks that are 150×150, and they’re almost all regularly used, not abandoned to gangsters and garbage dumpsters. So clearly the modal split and ped traffic supports it. But if you head to Midtown, there is NOTHING cutting up any of those blocks. Blocks that are 1000 feet on a side, in some of the most pedestrian-dense areas of the country. And there’s, what, Rockefeller Plaza? Nothing else?
To analyze NYC in a vacuum, it makes sense. All of the major transport is and has always been on the avenues. All of the major ped flows are towards and away from those avenues. And 200′ of width is wide enough for most buildings – or at least too narrow to leave enough room to cut up a single-deep row of lots. Olmsted actually complained about this in some of his writings, that the NYC grid forced tenement construction by leaving no blocks suitably small enough for workingmen’s rowhouses.
Maybe you could add a function that allows for longer lengths with shorter widths. But really, I don’t want to. Because between the Green Book and the HCM and Trip Generation, there are really enough publications out there that oversimplify complex and spatial issues into easy equations. Sometimes you just need to think qualitatively and visually, to leave a pattern as just a pattern.
Of course, if one of y’all wants to, by all means post the regression in the comments. 
keephoustonhouston
