WIDER SETTING AND FOOD FLOWS

Following consultation with colleagues at Makerere University and with the Mayor of Kawempe Division (1·2 million people), I researched in Bwaise-2 and Kawempe-1. In part a congested slum, Bwaise-2 ward has over 17,000 residents and is subject to periodic flooding; agriculture is hardly practised. Here I researched in Jambula zone, where ‘real slum’ conditions exist, and in Lufula, which is popularly said to have ‘more decency’ and a mixture of fortunes. Kawempe Division also has wards that do practise agriculture, as in Kawempe-1, whose population is double that of Bwaise-2. Urban agriculture in Kampala has a long history (see below), but its appeal gained momentum during the 1980s liberation war, when ‘the purchasing power of formal wages … was less than one tenth what it had been in 1972!’ (Maxwell & Zziwa Reference Maxwell and Zziwa1992: 10).

Historically, Kawempe Division, initially termed Kawempe Township, was outside Kampala's municipal boundary but within Greater Kampala (Southall and Gutkind Reference Southall and Gutkind1957: 12). Gazetted in 1955, the township ‘sprung out in chaotic spontaneity’, experiencing conditions similar to those found in other suburban pockets: among them, dense settlement, a lack of infrastructure and poor sanitation (1957: 17). Lying outside the colonial capital's boundary, the dominant form of tenure in the township was privately owned mailo land (Fortt Reference Fortt, Richards, Sturrock and Fortt1973: 68–9).

Following land reform and a succession of new laws early on in the 20th century, land in Buganda ceased to be controlled from the political centre; some estates remained in royal and chiefly hands, but vast amounts became available as mailo land to the Baganda population. This gave Baganda women ‘the legal capacity to inherit, purchase and own freehold land in their own right’. Baganda women now had a choice not shared by other women in the region: ‘they could either do what their families would expect of them and get married, or they could remain single and become economically independent land-owners’ (Ogden Reference Ogden, Werbner and Ranger1996: 172). The number of mailo owners, just under 4,000 in 1902, rose to 100,000 by 1965 (1966: 26). The trend slowed down, however, with the rising value of land in the 1960s (Richards Reference Richards1966: 26). Fathers now began to use their ‘testamentary freedom’ to leave their mailo land to whoever they pleased (1966: 25), thus turning siblings into potential co-heirs and rivals. Today, exploiting a family plot (kibanja) at home, which many Kampala residents do, indicates they have either inherited the plot or expect to do so.

Importantly too, where mailo land is subdivided and let, tenants have significant rights, and cannot easily be evicted. Historically, this has caused frustration among mailo landowners, especially in densely populated areas (Fortt Reference Fortt, Richards, Sturrock and Fortt1973: 83). This reality lives on in the 1998 land law, which stipulates that the mailo landowner must leave the tenant undisturbed unless s/he defaults on the rent (Hanson Reference Hanson2003: 239). The upshot in suburban Kampala, where mailo land has acquired high economic value, is that tenancy holders may be cashing in (e.g. through building shops or lodging houses for rent) whereas owners are unable to capitalise on the land. Southall & Gutkind (Reference Southall and Gutkind1957: 35) noted the anomaly already in the mid-1950s. The contemporary situation is complex. Where my respondents stated that they own the land they live on, which was the case in 58·5% of household heads interviewed, they may either hold the land title in their name or hold it in the form of a kibanja, which means that the long-term use rights have been purchased from the person holding the title (see also Maxwell & Zziwa 1999: 36).Footnote ¹

At the turn of the last century, suburban Kampala wards (previously called parishes) were known to have grown

When Kampala first attracted large numbers of migrants, their objectives and conditions were not conducive to setting up large families. Men mostly wanted to make some capital to invest on returning home (Elkan Reference Elkan1960: 42–3); women also aspired to making their own money, but many arrived as single women hoping to achieve independence and settle in town. They escaped ‘not only … from tribal sanctions, kinship obligations and male dominance, but in many cases from marriage itself’ (Southall Reference Southall and Southall1961: 224). In Namuwongo-Wabigalo ward, Christine Obbo learned that

Even when they had husbands, women who worked as petty traders normally worked on their own (Bantebya & McIntosh Reference Bantebya and McIntosh2006: 178). Invariably, however, urban women faced gender-related problems in their families and at work, leading to high divorce rates (Richards Reference Richards1966: 21).

For the wards I researched in, the 2002 Uganda Population and Housing CensusFootnote ² provides the statistics shown in Table I.

Table I.

Ward-level statistics

The food traders I conversed with, mostly in the markets of Kalerwe and Kawempe-1, told me that the city's food supply has undergone dramatic changes. Kampala used to get much of its food from the country's Central Region, which nowadays is said to suffer soil exhaustion. Today, produce comes from the Central, Eastern and Western regions, with the exception of matoke (plantain), which grows in abundance in districts west of the capital city: Mbarara, Rukungira, Masaka and others.

Matoke, rice, maize, beans, sweet potatoes and cassava are commonly consumed. Central to the Kampala diet, matoke is abundant from mid-January until the end of June, after which the availability goes down and the price increases steadily until it doubles by the end of the year. The period of scarcity occurs when local rice is unavailable (May–August) and more expensive rice is imported from Vietnam, Pakistan and Tanzania. The third key staple, maize, usually sold as flour, is sourced from all over Uganda. There is good availability from July until late February. For beans, the popular pulse, Buganda is the main provider from October until late April. The market price then rises until September, when bean mixtures (called mawanga magate, literally ‘mixed tribes’) are sourced from outside the district. Large-scale traders also create shortages and price hikes by hoarding local beans (called nabaale beans). Sweet potato, another important food, has a more complex cycle of abundance and scarcity, with July–August and January–February being times of low availability and high prices. Finally, cassava has a fairly regular supply, knows no significant price hikes, and is often eaten cooked and prepared as a stew (katogo of cassava). Fried cassava is a popular breakfast.

The Government of Uganda does not regulate food prices, but addresses ‘micronutrient malnutrition by conducting iron and Vitamin A supplementation for children 6–59 months and for women’ (Harvest Plus 2012: 1). Food diversification through nutrition education is also promoted. Food prices, however, are determined by the market in response to supply fluctuations. Low supplies are caused by seasonality, erratic weather conditions and the ever-rising cost of fuel for transportation. Recent years have seen heavy inflation in the price of food, with the exception of 2010 when there were bumper harvests for maize and rice (EPRC 2012).

When periods of scarcity – and thus high prices – are put next to each other, it becomes clear that July–August is a time when hardship sets in for the urban poor (see Figure 1).

Figure 1.

Food price fluctuations, Kampala 2012

Price fluctuations also reflect regional factors. Of importance here is Uganda's role as a key contributor to food security in East Africa. Vast quantities of beans and maize are exported to Kenya, Tanzania, Rwanda, the DRC and South Sudan. Maize exports can be informal or formal; in the latter case, a significant amount is funnelled through the WFP programme, which ‘puts direct upward pressure on maize prices inside Uganda’ (Gautam & Zorya Reference Gautam and Zorya2008: 6). This upward pressure may be good news for large-scale farmers, but not for vulnerable consumer groups like the urban poor (2008: 10). In September 2010, WFP staff informed me that 25% of Uganda's grain export was organised through their programme (staff 2010 int.).

Seasonal difficulties in food procurement also relate to the school calendar. In Kalerwe market, a dry-food trader made the following connection:

Consumers, too, linked the school calendar to procurement difficulties, though from a different angle. As Rose (pseudonym) told me,

Rose's mother, who was divorced and lived in, met the household's school fees: for her daughter's three sons (aged 18, 15 and 7), one nephew (aged 14) and one classificatory grandchild (aged 6). The fathers of these children, I was told, contributed very little.

HUNGER, SEASONALITY AND MEALS

I take people's perception of their food intake and its quality as my point of departure because this approach complements the work on metric measurements and nutritional value by health specialists. Also, global definitions of food security underscore the institutional and political aspects of an adequate diet; variables that cannot be quantified. Of the key dimensions to food security (availability, access and adequacy), matters of accessibility cannot be grasped when power relations and cultural expectations are left out (FAO 1996: 3). Targeted interventions, for example through nutrition education, may fail when local experiences of vulnerability are left unaddressed (Wheeler Reference Wheeler and Wood1985). As will be seen for the research area, coping with vulnerability meant achieving two meals a day and some diversity of ‘food’ and relish during the week.

At the onset of my research, in July 2012, matoke was still relatively abundant and affordable, except in the poorest households. Two months later – matoke now scarcer and growing in cost – more households had reduced their matoke intake; many were already pegging down to a late breakfast and just one meal a day. Hunger was said to have set in. That July marks a turning point for poor households was made explicit when an informant told me: ‘July is difficult because matoke, rice and sweet potatoes all become expensive. From July onwards, we eat one meal a day’ (26.7.2012 int.).

The period of research (July to mid-October) was marked by a gradual change in food availability, cost and intake. The interviews gave me a sense of what this change entailed, and when and why it came about. Unlike in Lilongwe, Malawi, where the perception of seasonal dietary change revolved around the low availability of (exceedingly expensive) maize meal towards the end of the calendar year, seasonal change in Kampala had to do with the availability and cost of matoke, which, when expensive, is initially replaced with imported rice (if rice can be afforded) and ultimately with a stiff porridge made from maize meal (posho). When rice can no longer be afforded, poor households switch to posho with beans in a groundnut sauce, and eat just once a day. The arrival of cheap(er) local beans in October is then eagerly awaited. In times of food stress, the relish should still contain some onions and tomatoes, and a leafy green. Popular greens are cabbage, spinach and kale; less common are indigenous greens like nakati (Solanum ethiopicum), ddodo (amaranth), and buga (Amaranthus dubious). Kale is locally known as ‘sukuma weekly’, meaning: eat once and it will push you for a week. Poor families do not eat vegetables daily.Footnote ⁴

For the urban poor the year has two halvesFootnote ⁵:

When matoke is abundant and cheap, low-income households may manage a meat or fish stew at the weekend. Silver fish (mukene) is particularly popular, as is offal when ‘meat proper’ cannot be afforded.

As already noted, the seasonal shift in staple food is gradual. When I interviewed her in September, Josephine, who did not consider her household to be food-insecure, reflected:

Mixing reduced amounts of matoke with rice was common. Ultimately, though, the poorer a household the more quickly it switched to posho with a simple groundnut sauce, and lived on one meal a day. Beans would be added to the relish as soon as cheap local beans became available (see Figure 1). The one meal a day turned into a late lunch taken around 4 pm – ‘Uganda time’! This late lunch was had when day-school children came home. If the parents or carers could afford to pay for school lunches, the children would already have had the standard posho-and-beans school lunch. In theory, the relish for the late lunch is not affected by seasonal change. But, as the hardship deepens, it will no longer contain fish or meat, and may have fewer tomatoes and onions.

The observed strategies of cutting down on food portion size, switching to less preferred foods, and – in the case of acute poverty – settling for just one meal a day, were short-term coping measures that were also recorded in the early 1990s (Maxwell Reference Maxwell1996). Respondents borrowed food and cash, as an interim measure before deciding that the actual number of meals needed cutting. Men in their early twenties said they often got ‘cash credits’ from a friend, perhaps two or three times a month, and that the amounts borrowed needed to be repaid quickly. Older informants, on the other hand, would first contact their grown-up children for assistance or approach a friend, or would obtain food on credit from a helpful shopkeeper or market vendor. In its assessment of how Ugandans coped with high food prices in 2008, the World Food Programme reported similar findings (WFP 2009: 79).

PERCEIVED INSECURITY, HOUSEHOLD STRUCTURE AND COPING MECHANISMS

In terms of nutritional value, the shift from matoke- to posho-based meals does not mean that the diet worsens. After all, as is well established, the mainly carbohydrate preferred diet of poor Baganda lacks protein, becomes unbalanced when children are weaned, and may lead to kwashiorkor and stunting (Welbourn Reference Welbourn1955a: 170). Also noting that plantains are deficient in protein, Elkan qualified the standard Ganda diet as ‘a diet which satisfies hunger’; it compared unfavourably with diets structured around millet varieties and root crops (Elkan Reference Elkan1960: 12–13). The diet of Kampala Luo, for instance, which was more cereal-based and varied, was found to be nutritionally far superior, as was that of middle class Baganda (Welbourn Reference Welbourn1955b: 294; Rutishauer Reference Rutishauer1965: 567). While shifting to a posho-based diet is not in itself detrimental to their health, what mattered to my informants was that the shift was virtually synonymous with pegging down to a single meal a day, without any improvement in the amount or quality of the relish they consumed.

Out of my sample of 118 households, 24 respondents mentioned having had just one meal on the day before the interview. They indicated that this one-meal-a-day had become the norm. Of those 24, 16 were grandparents who cared for grandchildren and sometimes nieces/nephews. The other eight food-insecure households involved adults whose income derived from poorly paid informal activity, such as peddling small non-food items, beer brewing, the sale of liquor, sex work, doing laundry, repairing shoes and selling street foods. Despite their reliance on poorly paid activities, the heads of poor households often cared for children other than their biological offspring. Quite a number of these children were orphaned, mostly because the parents had died of AIDS. However, it was often the sheer number of dependents, who frequently included unemployed ‘own children’, which plunged households into food insecurity.

The high number of grandchildren who were looked after by grandparents, mostly single grandmothers (26/38), was a key feature of the research scene. A small majority of these grandparents (22/38) thought their households stayed food secure throughout the year. For each of the four research zones, Table II lists the number of households in which grandchildren and/or nieces and nephews were being cared for.

Table II.

Caring for grandchildren and nieces/nephews

Roughly 40% (47/118) of the households surveyed had grandchildren and/or nieces and nephews in residence. This stands in marked contrast to the era around Uganda's independence when Kampala households had remained small. As Southall & Gutkind (Reference Southall and Gutkind1957) observed in Mulago, where ‘the family unit’ seemed the norm, its most striking feature was ‘the relative scarcity of children under five’ (only 9% of the population), a fact linked to ‘the custom of [parents] having their young children looked after by relatives who do not necessarily live in town’ (1957: 132). While it was true that the prospect of an advanced education was a magnet that brought rural children to Kampala, as Richards showed for peri-urban Kisozi (Richards Reference Richards1966: 66), the movement of children was by no means unidirectional.

How did the urban poor get by? Here is the telling case of Mary, a widowed grandmother who just about managed to keep her complex household food secure. At the time of our interview, she still managed to serve up two meals a day, with some diversity.

A NEW DISCOURSE ON CHILD FOSTERING

Child fostering among Baganda has a long history. According to Ainsworth (Reference Ainsworth1967), the practice of mother-child separation was traditionally sanctioned at the time of weaning, which frequently coincided with the birth of a new baby. The weaning child would normally be sent to a ‘grandmother, either maternal or paternal, who was chosen by the parents’ (1967: 417, emphasis added). A second reason for fostering – in the ‘old days’ – related to the recruitment of children to be brought up at the Kabaka's court. As a strategy for escaping such recruitment, it became ‘desirable to send [the] child to live with the maternal grandmother or another of the mother's relatives’ (1967: 417, emphasis added). A stay with maternal relatives, who were not of the child's own clan, provided safety when the recruitment officer called.

With relevance to the situation in Kampala today, Ainsworth added:

When Welbourn researched the possible benefits of fostering to Baganda children of weaning age, she argued that conditions of malnutrition might clear as a result of the fostering, but this was by no means a common outcome (Welbourn Reference Welbourn1963: 20). Ainsworth, for example, noted that it was ‘more frequently [observed that] the separation from the mother made matters worse rather than better. Too often the child was moved to a household where he belonged to no one in particular’ (Ainsworth Reference Ainsworth1967: 427). In actual fact, children fostered in rural areas might be reclaimed by their urban mothers precisely because they had developed kwashiorkor (Welbourn Reference Welbourn1963: 20). Given these mixed results, it was perhaps inevitable that child-mother separation for Kampala children had declined by the early 1960s (1963: 16).

Today there are signs that HIV/AIDS has rekindled the practice of child fostering, albeit with a thoroughly altered discourse. Looking after orphans has resulted in large three-generational households in which the middle generation is mostly missing. In my sample of 118 households, the average of size for households where grandchildren were fostered was 7·9 members, but about a quarter of them had 10 members or more. The explanation for such large households is that the AIDS pandemic makes ‘women, especially those with cash resources, commonly assume responsibility for orphans’ (Bantebya & McIntosh Reference Bantebya and McIntosh2006: 256). To this must be added the already noted ‘tradition’ that grandmothers cannot refuse to take in grandchildren (Ainsworth Reference Ainsworth1967: 417) – a practice which also appears to be growing.

The high incidence of grandparents who manage large, complex households has given rise to a discourse on ‘real grandchildren’. Several times in my research, grandmothers reflected on which grandchildren they would ‘ideally’ look after. Visiting Ruth, I asked her if it was true – as I had heard – that maternal grandparents were nowadays more likely to take care of grandchildren. She gave a thoughtful response: ‘When a daughter-in-law misbehaves towards the husband's side, the husband's parents may end up not loving her children. They will question whether their son really is the father of those grandchildren whose mother is misbehaving’ (4 October 2012 int.).

I took this up with another grandmother, Prudence, whose husband ran a successful small business. The couple's four children all had children of their own: one daughter was unmarried; one daughter had left her first husband; the two sons were divorced. The sons' children were staying with their biological mothers or mothers' mothers; the children of the daughters (aged 10 and under) were staying with Prudence and her husband. As I felt at home with this family, I put my question to Prudence.

Another informant endorsed this, adding that maternal grandparents care more ‘even when paternity is not in doubt’. She injected nuance: ‘Grandparents on the paternal side do not care much; they do not treat the grandchildren the way maternal grandparents do. The latter treat the children well; they truly care about them. But on the paternal side, grandparents only care if the father, that is their own son, is still alive’ (4.10.2012 int.).

The underlying logic is that when grandchildren are taken in, it is ‘better’ (more satisfying, less trouble) to look after a daughter's offspring. My kinship diagrams confirm this: children living with grandparents frequently belonged to daughters who had gone through a divorce and were starting new relationships. Sarah, another informant, was explicit on this point (28.9.2012 int.). Her household consisted of herself and the children who issued from the failed, first marriages of three of her daughters. Each daughter had seen her children rejected by the man with whom a new relationship was begun. So the children ended up staying with their maternal grandmother.

The growing ‘preference’ for looking after maternal grandchildren departs from practices recorded in the 1990s, but reconnects with the much earlier practice whereby maternal grandmothers fostered children lest they be recruited by the Kabaka (above). For Kampala in the early 1990s, Obbo (Reference Obbo, Jones and Nelson1999) showed that it was ‘tradition’ to call on paternal grandparents should grandchildren need to be taken care of. One particularly striking example she discussed was that of Mika, a young man who died of AIDS after fathering seven children with different girlfriends. On his death, and with the irate former girlfriends fearing for their own lives, all ‘seven children were “brought home to their father as custom demanded”’ (1999: 156). Describing the situation in Luwero, Guest wrote likewise: ‘Children belong to the father's family, according to customary law, so usually go to them after his death. Mothers are expected to go away and start a new family’ (Reference Guest2001: 98). In the 1990s it was commonly said that ‘when a woman does not produce, her husband's clan cannot grow’ (Ogden Reference Ogden, Werbner and Ranger1996: 177). In Kampala's poorer neighbourhoods today, the perception that a woman produces for her husband's clan, has lost much of its prominence – even though, in the absence of formal ceremonies, the birth of children continues to validate urban marriage (1996: 178).

ROLE OF RURAL FAMILY FARMS IN URBAN FOOD SECURITY

Farming ‘at home’ matters to the Kampala residents. Many women and men I interviewed, in fact nearly half of my respondents, had access to a kibanja under the mailo system. This greatly enhanced their food security. But there was a complicating factor: since kibanja is not formally divided, siblings must negotiate the terms on which the inherited land is to be allocated and used, especially when the deceased owner has left no will. As negotiations often fail to satisfy every party, informants may give vent to their frustration. Unhappy because she was not getting any produce from the family farm that her brother and sister cultivated, one disgruntled interviewee said of the hunger she knew in the city: ‘Food shortage starts at the village’ (bula lyemere litandi kira ku kyalo) (12.7.2012 int.). Farming or not farming a piece of rural family land made a big difference in the quest for urban food security.

I collected systematic information on the rural connections that respondents maintained with their areas of origin. I was interested to know whether respondents cultivated land at home and, if not farming, whether relatives provided them with food gifts, either occasionally or through the year. The results are shown in Table III.

Table III.

Rural contributions to urban food supply (114 households)Footnote ¹²

Households that actively farmed at home constituted almost half of my entire survey (54/114); with just three exceptions, all thought of themselves as reasonably food secure. Of the food-insecure households, half (12/24) did not farm in the areas of origin, nor did they receive assistance from rural kin. The remaining 12 were scattered across the other categories: five received occasional food gifts; three received food gifts more regularly; one farmed a plot ‘too small to yield anything’; and three both farmed and received produce year-round, but it was not enough. These last three were large, complex households.

The finding that rural farming makes an important contribution to urban food security in Kampala, at least in the wards where I researched, is particularly significant since so much of the literature on urban food security in East Africa focuses on the benefits of urban agriculture, leaving rural farming out of the equation. This can be deduced from a recent review by Diana Lee-Smith (Reference Lee-Smith2010). For Kampala, Lee-Smith looks back to

Because of its focus on examining the relevance of urban agriculture for urban food security, the study by Sebastian and colleagues ‘over-sampled’ farming households (Sebastian et al. Reference Sebastian, Lubowa, Yeudall, Cole, Ibrahim, Cole, Lee-Smith and Nasinyama2008: 71) and omitted to ask respondents about their interests in rural farming.Footnote ⁸ While this bias is understandable given the focus of enquiry, findings inevitably obscured how respondents may have simultaneously maintained a kibanja plot ‘back home’.

In my research, women regularly underscored the importance of having access to ancestral land. When caregiver Ruth (above), who headed a particularly complex household, told me about her coping strategies, she mentioned that a stepson, a son of her (deceased) husband's first marriage, was now the main source of financial support. He worked as a shop assistant in central Kampala. Other significant support came from Ruth's two brothers, who regularly sent staple food – matoke and cassava – from the 10-acre family farm in the area of origin. She said she was fortunate to have brothers who honoured her share of the family land (21.9.2012, int.).

URBAN AGRICULTURE IN KAMPALA

From early on, the mailo system has boosted urban agriculture in Kampala. Describing Kisozi in the 1950s as ‘typical of the peri-urban communities of Buganda’, Audrey Richards found the house-site plots ‘were small, averaging 3 acres, a figure well below that for Buganda as a whole, but not for the densely populated areas round Kampala’ (Richards Reference Richards1966: 67). Peri-urban agriculture also yielded food cash crops, as Obbo (Reference Obbo1980) learned during fieldwork in Namuwongo-Wabigalo. Whether they had been born in the area (as daughters of migrants) or had arrived there to settle, several women in her study supported themselves by earning money from cultivating land they had inherited or purchased (1980: 71–2). As a general rule, mailo holdings remained sizable and productive the farther removed they were from the city centre (Elkan Reference Elkan1960: 46).Footnote ⁹

Kampala is often praised for being at the forefront of the urban farming campaign. It has long been appreciated that ‘Kampala City Council is … unusual, not only in having a Department of Agriculture but also in having developed both by-laws governing urban agriculture and a system of classifying types of urban agriculture zones’ (Lee-Smith Reference Lee-Smith2010: 486). This unusually positive situation has galvanised urban and peri-urban agriculture. However, and this chimes with my own observations, Lee-Smith argues for Kampala and other African cities that ‘better-off households are currently benefiting more from urban agriculture than the majority of poor households' (2010: 483).

In the two wards where I researched, agricultural practices differed markedly. In crowded Bwaise-2, no one cultivated around the house (lack of space) or on rented plots (risk of flooding), and there was only limited interest in raising small livestock. In contrast, the enthusiasm for urban agriculture in Kawempe-1 ward was robust. In the ward's Kalule zone (30 households surveyed), 10 had sizeable gardens which produced both crops and fruits, for home consumption and sale. Tall fruit trees and extensive matoke groves were a common sight. Besides matoke, the selection of crops included maize, beans, cassava and sugar cane. Entrepreneurial farmers, who enjoyed relative wealth, might also specialise in marketing a range of vegetables, including mushrooms. Some of their businesses generated local employment. In Kalule, 7/30 respondents raised between 150 and 300 chickens each, while one extended household ran a chicken farm with over 2,000 birds. A further two households raised 20 chicks each. Some households kept goats (half a dozen on average); one raised two heads of cattle; another kept healthy looking pigs. Among the households that did not cultivate around the homestead, three rented a plot elsewhere in Kawempe-1, where they raised crops and kept small livestock. This gave a total of 13/30 households which actively practised urban cultivation, and a total of 9/30 who raised poultry. The scenario in Kalule, which seemed replicated in Kilokole zone (where I surveyed only 12 households), confirmed the strong interest in urban agriculture reported for other parts of the city (see Sebastian et al. Reference Sebastian, Lubowa, Yeudall, Cole, Ibrahim, Cole, Lee-Smith and Nasinyama2008). But there was a caveat. Although Kawempe-1 undoubtedly exemplified the potential for urban agriculture, there was continuing pressure on households to convert ‘farmland’ into building plots for rentals. If one had the land, constructing rentals was the easier route to generating some regular income.Footnote ¹⁰

The information I gathered in Kawempe-1 ward conforms with recent observations from Eldoret, Kenya, where urban residents have taken up ‘livestock keeping mostly as a means of earning additional income but also for additional food’ (Simiyu & Foeken Reference Simiyu and Foeken2013: 598). Significantly, despite Eldoret's stringent legal and policy restrictions, livestock keeping had become common practice: ‘of the 160 urban households surveyed, 73% kept livestock – 60% kept livestock in addition to cultivating crops, while 13% kept livestock only’ (2013: 584). For Kampala, moreover, the literature suggests that agriculture's dual purpose – as a source of additional income and food – has been in evidence since well before Uganda's independence.