Urban farming in Italy — the 59 farmsteads a city owns, and the 41 meadows it counted

When we talk about preserving urban farmland we tend to think in areas: how many hectares were saved, what share stayed in cultivation. The roughly 46,300 hectares of the Parco Agricolo Sud Milano, which yesterday's instalment described, is usually recounted in exactly that register. Yet stand inside the line and the landscape refuses to be explained by area. What is there is a brick farmstead built as a square around its courtyard; a spring where groundwater surfaces of its own accord; a meadow engineered to stay green through the winter. Each is a particular thing, and each has a name. The decisive fact for understanding Italian urban agriculture is that preservation here has always taken objects, not areas, as its subject.

The clearest expression of this is the Milanese cascina. As the Italian Wikipedia entry sums it up, the cascine still surviving in Milan number 'un centinaio' — about a hundred — and 'cinquantanove delle quali fanno parte del demanio comunale': fifty-nine of them belong to the municipal estate. The role Milan plays here is therefore double. It is the regulator that decides, through planning, how farmland may be used; and it is the landlord that owns the farm buildings themselves. The Southern European technique yesterday's article described — holding land without buying it — turns out, in Milan's case, to have been only half the story.

It feels intuitive that owning something is how you protect it, but that is not quite right. The same source goes on to record that, of the hundred-odd cascine, 'diciotto sono diroccate' — eighteen have fallen in. Even with a public body that intends to preserve them holding the title, close to one building in five has become a ruin. This is less a matter of negligence than of what ownership is as an instrument. A boundary, once drawn, costs almost nothing to maintain; owning a building entails an outlay every year — roof, beams, drainage, and, where the structure is listed, the whole protocol of proper restoration. The concreteness that is the strength of the inventory approach is also, precisely, the concreteness of the cost.

What follows first traces why the cascina, as a form of building, survived inside the city at all, and looks at the thirteen of them still being farmed by tenants. It then establishes that the city is not the only landlord — a hospital, a public housing agency and the diocesan curia each hold farmland too — before turning to the most refined version of the inventory method, the 41 marcite that the southern agricultural park has counted and placed under protection. Finally it considers what it means that, under this method, failure takes the shape not of a cleared lot but of a ruin.

The cascina is a farmstead type particular to the Po plain. Farmhouse, byre, barn and workshop — often workers' dwellings and a chapel as well — are laid out as a rectangle around a courtyard. A single cascina is less one farmer's house than something close to a small hamlet. The form arose because agriculture here was intensive and combined water management, livestock and grain, which required many hands and many buildings gathered in one place. The oldest in Milan is held to be Cascina Chiesa Rossa, which the Italian Wikipedia entry says 'risale al X secolo' — dates back to the tenth century. These stood out on the plain from a time when the city itself still fitted inside its walls.

Why, then, is a farmstead of the open plain located 'inside Milan'? The answer lies not in agriculture but in the history of administrative boundaries. Milan absorbed the independent municipalities around it in stages. The same source gives the case of 1923: 'Nel 1923, anche il comune di Baggio, ultimo Sindaco Cesare Stovani, fu soppresso e annesso alla città di Milano' — the comune of Baggio was abolished that year and annexed to the city; the entry records the same for Lambrate. The cascine did not move into the city; the city's boundary moved out past them. The order matters. Much of Italy's urban farmland is not greenery that planning created, but agriculture left on the inside when the administrative line expanded over it.

This history sets the character of Italian urban farming. Agriculture that a city has enclosed after the fact was not designed to the city's logic. Field sizes, the arrangement of buildings, the direction the watercourses run — all follow a rationality that made sense when everything around was farmland. Trying to maintain that inside a city necessarily becomes the work of fitting together things that do not fit. What the municipality took on as landlord was not only the land but this mismatch itself.

Of the hundred-odd cascine, how many are actually farmed? The Italian Wikipedia entry is strikingly specific: 'Tredici, però, in piena area urbana, sono ancora condotte da fittavoli, secondo l'uso milanese, come aziende agricole in attività' — thirteen, in the middle of fully built-up territory, are still run by tenant farmers, in the Milanese manner, as working agricultural businesses. The phrase to notice is 'secondo l'uso milanese,' according to Milanese custom. Ownership sits with the city; the enterprise sits with the tenant. Keeping the party that holds the land separate from the party that tills it, and sustaining agriculture through a customary tenancy, is the mirror image of yesterday's 'holding without owning': here the city owns, and arranges not to farm.

What they produce is equally concrete. The same entry lists 'la cascina Battivacco alla Barona (riso), la cascina Basmetto alla Barona (riso), la cascina Campazzo (latte) e la cascina Nosedo (latte e derivati)'. Battivacco and Basmetto, in the Barona district, grow rice; Campazzo produces milk, Nosedo milk and dairy products. Inside the administrative territory of a city of more than 1.3 million people, flooded paddies are maintained and dairy cattle are kept. Rice survives here because of the water regime. The irrigation system of the Po plain was built up over centuries with rice in mind, and the shape of the channels and fields is optimised for it. Yesterday's instalment argued that water is the first constraint on Southern European urban farming; in Lombardy the constraint points the other way. Water is not scarce but close to abundant, and it is abundance that has dictated the form of farming.

Thirteen can be read as many or as few. Against a hundred it is a small share. Yet the fact that commercial rice growing and dairying persist on land a few kilometres from the centre is exceptional for a major European city. What matters is that these thirteen are not accidental survivors but are held in place by an institution — the tenancy. As long as the arrangement holds, the thirteen hold; if it lapses, if no tenant can be found, the number falls. What the inventory method preserves is the building and, just as much, the relationship through which the building is used.

If 59 are municipal, who holds the remaining forty-odd? The Italian Wikipedia entry answers: 'Gli altri proprietari sono l'Ospedale Maggiore, l'Aler (ex Istituto Autonomo Case Popolari), la curia' — the Ospedale Maggiore, Milan's principal hospital; Aler, the public housing agency formerly known as the autonomous institute for working-class housing; and the diocesan curia. Private property companies make up the rest. The city's farmland is divided among a hospital, a social-housing body, the Catholic Church and real-estate firms. To a Japanese reader this may look strange; it is in fact a common residue of European urban history.

Hospitals and religious orders hold large tracts of farmland because donations and bequests to charitable institutions accumulated over centuries. A hospital met the cost of treating the poor out of the income from donated land, and that land was held as an asset across generations. When the financing of medicine shifted to insurance and taxation in the modern era, the land alone remained. A housing agency comes to hold farmland by a similar route: of the sites acquired for construction, the parts never built on stay on the books as fields. These holdings, in other words, were not kept for the sake of agriculture. They were acquired for some other purpose and simply never disposed of once that purpose lapsed.

This dispersal of ownership cuts both ways for preservation. On one side it is a strength: with many owners, the risk that a single fiscal crisis or change of administration sells everything at once is low. On the other it is a plain weakness. To maintain one agricultural landscape you must align the intentions of the city, a hospital, a housing agency, the curia and several property companies. This is exactly where the agricultural park, the device described yesterday, earns its keep. A boundary binds regardless of who owns what; it fixes the permitted use across a whole zone without any negotiation over title. Italy has both the inventory and the line, but they do different work. The line aligns use across dispersed ownership; the inventory protects particular objects by name.

The inventory method appears in its most refined form in the protection of the marcite. A marcita is a form of irrigated meadow peculiar to the Po plain: water from a spring known as a fontanile is kept flowing thinly and continuously through the winter, preventing the soil from freezing and keeping the grass growing. The key variable is not the volume of water but its temperature. According to the Italian Wikipedia entry, water from the fontanili emerges at a nearly constant temperature between 9°C in winter and 14°C in summer. Groundwater varies far less by season than the air, so in winter it is warmer than its surroundings. Running that warmer water thinly and without pause produces a green meadow on a plain where snow falls — a technique that uses water as a medium for carrying heat.

The payoff shows up as yield. The same entry records that the method allows at least seven cuts of forage a year, and often nine, against the four or five obtained from even the best permanent meadow — close to double. Keeping livestock near a city demands a great deal of fodder, and a meadow that can be cut nine times a year is what supplied it. Part of the reason dairying took hold around Milan lies here. The inventor is unknown, though Cistercian monks from France are generally credited with doing much to spread the technique through northern Italy: the engineering and agronomy accumulated in medieval monasteries laid the foundation of a city's food supply.

And the manner of protection is this article's subject exactly. Of the southern agricultural park, the Italian Wikipedia entry records: 'Nel parco sono state contate 41 marcite (punti dove i campi agricoli vengono irrigati con l'acqua delle fonti nel terreno), messe sotto protezione' — forty-one marcite were counted within the park and placed under protection. The passive 'sono state contate,' were counted, is the operative phrase. Not so many hectares protected, but a population of forty-one established first, and protected after. The same source notes that the marcite were abandoned through the twentieth century in favour of industrial maize and are now increasingly rare. To protect something that is disappearing you must first know how many of it there are; the logic of the inventory is contained in that sentence.

How finely this hydrology is tied to the terrain shows in the individual entries. The cascine source records that 'La ruota a pale del Mulino Dorino era mossa dalle acque del fontanile Cagnola, che confluisce nel fiume Olona a Pero' — the paddle wheel of the Mulino Dorino was turned by water from the Cagnola spring, which joins the river Olona at Pero. A spring, a mill wheel, a river, a place name. Protecting subjects described at that resolution is an entirely different job from setting a target in hectares.

Return to what the inventory method costs. That 18 of the hundred are 'diroccate' — fallen in — states its failure rate plainly. But the shape of the failure is worth attending to. When farmland preservation fails in an American city, the land is typically sold, cleared and developed. In Milan it goes differently: the building stays, the roof comes down, and the walls decay standing up. The difference follows from the form of ownership. Public property cannot easily be sold, and a structure of historic value carries constraints on demolition. The result is a condition that can persist for a long time, in which the thing can be neither sold nor demolished nor repaired.

That condition is not simply bad. A ruin leaves more room for recovery than a cleared lot does: while the walls stand, restoration remains possible whenever funding appears. In practice a substantial number of Milan's cascine have been repurposed — as libraries, recreational facilities, care facilities for the elderly, disabled people and those recovering from dependency, and as residential centres. Accepting non-agricultural uses lets the building survive and leaves room for farming to return later; it can be read as a long-run version of meanwhile use. This is not, however, grounds for optimism. The figure of 18 is precisely the count of buildings for which that recovery has not yet arrived.

The practical implication carries over to Japanese urban farmland. The idea that bringing farmland or its buildings into public ownership will preserve them often accounts for the cost of acquisition while ignoring the cost of upkeep. Milan's ratio of 59 to 18 makes the point rather harshly: taking title is not the end of preservation but its beginning. Ownership does not buy the power to protect; it buys the obligation to keep protecting.

Milan also offers an example that runs the opposite way from preserving what remains: the Parco Nord Milano, spread across the city's northern edge. The Italian Wikipedia entry gives its area as 640 hectares, straddling seven municipalities — Milano, Bresso, Cusano Milanino, Cormano, Cinisello Balsamo, Novate Milanese and Sesto San Giovanni. What matters is where it came from. The entry records that it was 'realizzato inizialmente sulle aree dismesse, subito dopo la seconda guerra mondiale, della Breda Aeronautica' — created initially on the land left disused, right after the Second World War, by the Breda aircraft works. This was neither farmland nor green space to begin with. It was an aircraft factory.

The greenery did not return on its own; it was manufactured. The same entry states that 'I primi interventi di rimboschimento furono avviati a partire dal 1983' — the first reforestation works were begun in 1983 — carried out by the Lombardy regional forestry agency under the direction of Paolo Lassini. A public regional body planted a forest on a factory site over a span of decades. Lombardy had recognised the area as a regional park in 1975, eight years before the planting started: the institution came first and the substance caught up afterwards.

Set the southern agricultural park beside this northern one and the range of Italy's hand becomes visible. In the south, existing agriculture was kept by drawing a line around it. In the north, greenery was rebuilt on land that had lost it. The first is cheap and quick but useless where the subject no longer exists; the second applies to any site but takes decades and public money. Which you can choose depends on what is left on the ground. And counting what is left — about a hundred cascine, 59 of them municipal, 18 ruined, 13 still farmed, 41 marcite — is the work of the inventory.

Let me reconstruct the path this article took. The starting point was the Southern European technique set out yesterday in 'Farming a dry city': holding land without buying it. This piece stepped inside that line and found not area but particular objects. It first traced the cascina as a building type, from Cascina Chiesa Rossa in the tenth century to the annexation of Baggio in 1923, establishing the order in which the city swallowed the farmsteads. It then examined the thirteen working farms and the relationship by which tenants till municipally owned land 'in the Milanese manner'. It went on to show ownership dispersed among the city, a hospital, a housing agency, the curia and property companies, and explained why the device of the boundary earns its keep precisely because it binds across that dispersal. From there it moved to the 41 marcite, and closed by setting the 18 ruins against the opposite pole of the Parco Nord, a greenery that was manufactured.

Let me restate the three most important points by name. First, Milan is both the regulator of farmland and its landlord: 59 of the roughly 100 surviving cascine belong to the municipal estate, and that double position is what makes the Italian case distinctive. Second, ownership does not guarantee preservation. Of that same hundred, 18 are ruins and only 13 are working farms; the ratio of 59 to 18 shows that taking public title is the beginning of preservation and not its end. Third, Italian preservation is implemented as an inventory. The clearest instance is the Parco Agricolo Sud Milano protecting 41 marcite only after counting them — meadows where spring water emerging at a nearly constant 9°C in winter and 14°C in summer permits seven to nine cuts of forage a year, far above the four or five a normal meadow yields.

Now what is not known, and what should be held in reserve. The figures used here rest mainly on the compilations of the Italian Wikipedia. The numbers — about 100, 59, 18, 13, 41 — are encyclopedic tallies, not the results of an official municipal census, and no reference date is stated for them, so the present counts may well differ. I would also have liked to cite national statistics on Italian allotment gardens, the orti urbani, from the environmental survey ISTAT conducts across provincial capitals; the institute's site could not be reached from this run's environment, so it went unverified. This article therefore makes no claim at all about the national scale of plots allotted to citizens. The strongest objection available is that Milan is exceptional and not representative of Italy, and that is fair: the Milanese case cannot be generalised to the country. What has been shown is one pattern, in a city with a particular hydrology and a particular history of annexation.

Finally, where this sits in the series and where it hands over. Day one showed that Southern Europe has held its farmland without owning it. Today, through Milan, we have seen the underside of that: the city does own, does let, and does sometimes fail at upkeep. Owning and not owning are not competing strategies but a division of labour inside the same city. Tomorrow's closing instalment visits Spain. Where Italy answered the problem with ownership and an inventory, in Spain the crisis after 2008 applied a different pressure, and the constraint of allocating water appears in a harsher form. Where did a city tilled by crisis, on dry ground, end up going? That is how this short series will close.