Heather Ann Kaldeway on ‘The need for concentration’ — How much density is enough? How much is too much?


This chapter focuses on the fourth of Jane’s generators of diversity: “A sufficiently dense concentration of people, for whatever purpose they may be there. This includes people there because of residence.”

She recommends using a Goldilocks Principle for density: Not too low, not too high, but just right.

Goldilocks and the Three Bears by Flickr user Helena Perez García

You want residential density to be high enough to fill the neighbourhood with people who use the streets, parks, schools, offices, grocery stores, bookstores, community centres, theatres, comic book stores, bike shops, coffee houses, music venues, etc., particularly in the evenings and weekends. However, this concentration of residences must also be low enough that it can be diverse itself, including many different types of housing:

“City dwellings have to be intensive in their use of the land for reasons that go much deeper than the cost of land. On the other hand, this does not mean that everyone can or should be put in elevator apartment houses to live — or into any other one or two types of dwellings. That kind of solution kills diversity by obstructing it from another direction.”

So we need concentration to stimulate neighbourhood diversity, but if we only build tiny bachelor-size condos, we squelch diversity in another way.

Then she starts throwing some numbers around. Let’s be frank:

I struggle to visualize the numbers Jane uses when she talks about density.

Here we go:

“I should judge that numerically the escape from ‘in-between’ densities probably lies somewhere around the figure of 100 dwellings to an acre” (italics in original).

At this point, I felt a bit like this guy:

Photo by Flickr user Rob Baird

Does my neighbourhood meet that criteria? How can I tell? How many net acres are in my neighbourhood? And what about units? I can count balconies on the apartment tower down the road, but what about the rest? Does Toronto publish such data?

She mentions specific districts as examples: What did they look like in 1962? Would 2012’s Google Street View be illuminating at all? (I suspect not.)

So, like all good geeks, I started researching.

In my hunt for answers, I discovered the Lincoln Institute of Land Policy’s Visualizing Density site (and it’s fabulous). They have this to say about Measuring Density:

What is the density of the street you live on? The answer is not any one number but several, depending on how you measure density and how broad an area you include in your calculation. Density can be expressed in different ways — persons per square mile, units per acre, or floor area ratio. Residential density is typically expressed in housing units per acre and measured as net or gross.

But do we need to get technical about it? We all know what high density feels like, right? Let’s test your density-meter:

Which of these has a higher density?

Development in Boulder, CO from Visualizing Density Image Gallery
Development in Detroit, MI from Visualizing Density Image Gallery

Trick question! They have about the same density. Surprised? Let’s zoom out and get some more details:

Development in Boulder, CO from Visualizing Density Image Gallery
Development in Detroit, MI from Visualizing Density Image Gallery

Specifically, the Detroit development puts 325 units on 9.6 acres (33.9 units/acre) and the Boulder development puts 191 units on 5.8 acres (33 units/acre). The Detroit development has low ground coverage, while the Boulder development uses more of its land.

Get all the technical details about density levels, along with lots of great examples of density in action, at the Lincoln Institute of Land Policy’s Visualizing Density site.

Improve your Density-Meter

Try the What Does Density Look Like? quiz:


(You can watch the entire 50-minute presentation online.)

The Western Australian Planning Commission has a great document which I found very useful: Residential Density and Housing Examples (PDF). It illustrates different levels of density in Perth, Australia. They provide examples of housing ranging from 10 units per net acre all the way up to 190 units per net acre.

How does Jane measure density?

She focuses on districts, not individual lots.

This is important to bear in mind, because it enables variety. Since we’re measuring at the district level, we can mix intense developments like condo towers and apartment buildings with houses and low-rise developments and reach a fairly high density for the neighbourhood overall. Densities become too high when they “begin to repress diversity instead of to stimulate it.” Variety is vital!

No one way is a good way to house a city neighbourhood; no mere two or three ways are good. The more variations there can be, the better… It is not easy to reconcile high densities with great variety in buildings, yet it must be attempted.

She measures units per net acre (or residential acre).

If you measure by net acre, you only count the land zoned for housing. In a district with a lot of short blocks, more land will be taken up by streets and sidewalks, and that land isn’t counted and therefore doesn’t lower your density measurement. However, if you build a huge yard, that under-used land will get counted and lower your density measurement. This measure focuses on how well you’re using the land available for housing.

If you measure by gross acre, you count all the land, whether it can be used for housing or not. This measure focuses on how much housing is in the area in general.

No one else seems to measure density this way.

Perhaps I’ve been looking in the wrong places, but I can’t find anyone else who measures density by units per net acre at the district level. I’m not even sure how she managed it in the first place, since she doesn’t cite her sources. Any suggestions?

Toronto measures density by gross floor area. If you know how to translate that into units per acre, I’m all ears. I suspect it’s not possible, but I’d love to see if it is!


I’m still not entirely sure what Jane’s neighbourhood district with 100 dwellings per net acre looks like. However, I do feel more confident on the following points:

  • Measuring at the district level asks us to take the context into consideration when building on a specific lot. Is a tower appropriate here? It depends on the neighbourhood. It could be just what’s missing, or it could be squelching diversity.
  • Thinking in terms of net acres also informs how we treat non-residential land. If we have high ground coverage and high density on our residential land, we must consider: Are our streets wide and welcoming? Do they leave us room to breathe, or are we trapped in a glass canyon?


  • One thing we haven’t talked about is unit size. If I build a 10-storey building with one unit per floor, that’s pretty low density but would allow for large family units. If I pack my building with tiny bachelor units, then we have more units per acre. Does anyone control for this? Perhaps we could measure by number of bedrooms?
  • Jane’s lower limit for residential density is 100 units / acre. What’s her upper limit? She doesn’t offer one. She knew building technology was changing rapidly in 1962 and would continue to change, so she stuck with the Goldilocks Principle: high enough to enable diverse primary and secondary uses, and low enough that you can still have a mix of housing types. We’re able to build much higher now, so we can put more people in a neighbourhood while maintaining a diversity of types than was possible when she wrote Death & Life.
  • Evisioning infill: Urban Advantage has some intriguing illustrations of how we can infill our spaces to add more density.
  • Did you know that Toronto currently has the most high rise construction in North America? See Page 15, ‘Economic Dashboard’ (PDF), City of Toronto.
  • The vast majority of 132+ highrises built in Toronto in 2011 were condos. As reported in The Toronto Star in October 2011, “you could probably count on your hand” the number of rental high rises and office buildings going up in the city. Are we threatening our balance of primary mixed uses by building so many condos and so few buildings of other types?

9 Comments Add yours

  1. Eric O says:

    Heather, I’m with you! Part of the reason folks don’t understand this chapter readily is we need to really get a grip on what these numbers actually mean visually in dwelling unit per net acre terms (DUA for brevity). Moreover, we need to actually understand what multi-family buildings require in terms of land, in the first place. How buildings use land is the ingredient that many planners sadly seem to sometimes miss.

    Ed Glaser, for example, claims that in Jacobs’s ideal world, which he claims is in the ballpark of 150 DUA, we wouldn’t build above six stories. But 150 dwelling units per net acre does not look like 6 story buildings. The reality is that 6 stories can hardly get to 100 DUA! It wouldn’t even scrape 80 DUA if you actually had to provide some parking spaces.

    The pristinely abstract mind estimates 6 stories for 150 DUA based on mathematics: 1600 sf x 150 divided by an acre. Briefly, this reveals Glaeser has no idea what he is talking about. That’s because dwelling units also need common areas, oodles of equipment, and – very importantly! – access to daylight. Not to mention parking. 150 DUA does not look like a six story building at all. Actually, it is more like 8-11 stories in most of the ways that zoning allows to build apartment buildings today in the vast majority of North American cities. Strange that very few planners stopped to call Glaeser out on his simplistic mathematics. The kind of density most people associate with Jacobs is really VERY inadequate density for Jane Jacobs.

    Basically, your greatest barrier to get at diverse density is parking minimums. To get to 100 DUA plus, you cannot do it without structured parking. Having to build with structured parking automatically means you have to subsidize units to keep some of those affordable.

    Here’s the dirty secret: When most of us think of “medium density”, i.e. the kind of low or mid-rise urbanism that most mistake Jane Jacobs was referring to in these numbers, simply does not come even close to the 100 DUA minimum required to support streets lined with retail and urban dynamism. You are right, with Jacobs building height was not the important physical parameter to measure good density in. It was ground coverage that mattered to her, and it is that exactly which most affects our ability to get at density that high in net terms. But today we can’t achieve 100 DUA in most places, for an important reason: the need to integrate parking and sheer economics. Most of us build “urbanism” at what to Jacobs was “inbetween” densities (densities just above suburban density but to feeble to support urban life). Most egregiously, we call it HER density. But this density always needs something else to support its vitality, because it simply does not have the people. Not enough. And the diverse people it requires, according to her.

    The reality is that most, if not all urbanist projects today, except in the densest, transit-friendliest cores, are achieving district coverage at at least 100 DUA. You cannot get to 150 DUA today without building some very tall buildings, and plentiful parking structures, sprinkled in among those 4-6 story mid-rises.

    1. Hi Eric,

      You’re right, parking is an issue — depending on how you build it. Here in the downtown core of Toronto, we’re seeing a lot of very deep parking lots built below buildings, even short buildings. There’s a particular development I’m thinking of that is either 6 or 8 stories, but has at least that many underground stories of parking. It has small shops on the ground floor, a huge grocery store on the second floor, a huge clothing store on the third floor, and then condos on the remaining floors. Parking is provided for all of the stores and all of the condos, but the whole floorplate is used by the building since the parking is tucked underneath.

      Of course, we also have areas in the city where this isn’t possible simply because the ground is too difficult to dig into (e.g. rock) or not stable enough to support this type of development. Then you get into underground infrastructure (covered rivers, subways, telephone/internet connections, sewers) and it all gets complicated quickly! 🙂

  2. Eric O says:

    Ooops…I meant most urbanist projects are NOT achieving district coverage above 100 DUA. 🙂

  3. Matthew says:

    I just would like to point out that she does give an approximate upper bound: 200 dwellings per net acre. The North End’s 275 du per net acre is excused by virtue of its heritage. I agree that the “Goldilocks method” is more suitable and probably what she would prefer people to think about.

    1. Matthew, interesting point! I did think about that, but I didn’t think it was fair to go on about this at length by the end of my (longer than usual) post. But now I can! 🙂

      Here’s my logic:

      You’re right that she mentions a figure of 200 units per net acre. However, as you also mention, she describes the North End already breaking that guideline without causing problems, because it was providing this density without squelching diversity (due to its “peculiar and long heritage of different building types,” p.217 — not, I would note, due to “heritage” in today’s “pretty old buildings” sense of the word). So she is open to the idea that a neighbourhood can be much more dense than her suggestion, and still be a good, thriving, healthy neighbourhood even in terms of density and diversity.

      Her main question about the upper limit of density is this: “Just how high can a neighborhood’s densities go without sacrificing the neighborhood to standardization?” (p.216)

      On page 212 she writes that “a numerical answer means less than a functional answer (and unfortunately can even deafen the dogmatic to the truer and more subtle reports that come in from life).”

      On page 213 she notes that the height of elevator apartments is limited by how quickly an elevator can go from the top to the bottom and by how high we can pour reinforced concrete. She notes that “[this] depends on the technological improvement of cranes, so this figure increases every few years. As this is written, it is twenty-two stories.” (Here in Toronto 2012, we’re in the midst of building a 66-storey residential building.)

      On page 214 she’s talking about elevator apartments and says they “do not provide standardization by virtue of being elevator apartments … [but] when they are almost the only way a neighborhood is housed.”

      Given that her lower bound is due to concentration being too low to provide the population needed to support primary and secondary uses, I think she would stand by that number.

      However, her higher number (200, as you noted) is solely due to concerns about standardization. In a world where those elevator apartments can be three times as tall as was possible when she wrote this, it is now possible to have higher density and high diversity in building types.

      Given that, I doubt that she would stand by the suggestion of 200 units/acre (and I think she didn’t even in the book, by immediately describing a neighbourhood that didn’t fit this mold). Your thoughts?

      Page numbers from 1992 Vintage Books edition — the light yellow one

      1. Matthew says:

        I pretty much agree with everything you said. We are capable of more density with diversity nowadays. But I do wonder if there is an upper bound relating to the notion of “eyes on the street.” Tall buildings, no matter how varied, are going to be somewhat impersonal to the street. This can be dealt with to some extent. But if everyone is hiding up in their apartment and leaving the street empty for long periods, then there is a problem. Of course, this also falls under the heading of “lack of primary uses.”

      2. I would go so far as to say that the street-level experience of a building has nothing to do with its height. Tall buildings have just as much (if not more) opportunity to create vibrancy on the street-level.

        There is nothing about a tall building which prevents small retail and offices on the ground floor, and there is much room within a tall building to add mixed primary uses (retail, galleries, social services, offices, parking, housing). These additional uses would draw eyes to the street at different times of day and for different reasons.

        The fact that we waste this opportunity constantly has less to do with height than with economics. Mixed use requires more complex building management and many of those uses are leased rather than sold (e.g. office space), which is not attractive if you’re building something with the intention of selling it immediately.

        (There is no reason offices could not be sold as non-residential condos, but that’s another conversation.)

      3. Matthew says:

        You are right that we are absolutely capable of constructing tall buildings that interact well at street-level, but yet, for some reason it often doesn’t happen. Not only will it require more complex management but everything they build will be new and therefore, high overhead.

        With that in mind, here’s a fairly old building I spotted today in Brooklyn with lots going on at street level:


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s