The Reproduction of Urban Capitalism: Street Food and the Working Day in Colonial Mombasa

Abstract This article argues that street food was an essential part of the social reproduction of Mombasa's working class during the colonial period. Like in other expanding capitalist cities, as Mombasa grew, urban workers lived further from their place of employment, which meant they could not return home for their midday meal. Street-food vendors provided them lunch at low prices in convenient locations, and therefore reproduced the working day by provisioning the calories that bridged morning to afternoon. However, postwar municipal authorities also wanted to create a particular kind of urban society in which the ‘informal’ activities of street-food vendors did not fit, and tried to expel them from the city's streets. As these campaigns unfolded, an unresolved contradiction emerged between this elite view of Mombasa, and the reality that the services vendors provided were necessary for the reproduction of the city's economy.

eateries had become indispensable to the reproduction of the working day in East Africa's most important port town.
Historians have explained rather well how workers in colonial African cities were able to purchase raw ingredients such as milled grain, vegetables, and other such items at markets and shops, but we know far less about how urban food was prepared. 5 In Mombasa, much of this kitchen work was done commercially by street-food vendors, running a type of business that also played an essential role in the global history of urban capitalism. Since 1800, the rapid growth of cities around the world has transformed the labor process of cooking. For most of human history, kitchen labor was done primarily as unpaid domestic work by women inside their homes, but proletarianization created fast-growing urban working classes who needed to pay someone to cook their food, precipitating the mass expansion of eateries of various kinds. 6 This was driven not simply by the general commodification of economic life under capitalism, but also by the way modern urbanism engendered new spatial relationships between home and work. Scholars of industrializing societies in the West have noted how the growing geographical size of cities made purchasing a midday meal a new necessity for workers. Indeed, during the early phases of industrial urbanization, people often lived close to work and returned home for lunch or had family members bring them cooked food. The continuing growth of cities, though, increased the distance between residence and worksite to the extent that accessing food from home during a lunch break became challenging or nearly impossible. 7 To sell their labor during a long working day, whether in New York, Chicago, Mumbai, or indeed Mombasa, people needed to pay someone else to prepare their lunchtime meal. In short, affordable and spatially accessible eateries became a vital part of the social reproduction of modern urban societies. 8 Since the 1980s, scholars have produced numerous studies on the social reproduction of laboring classes in urban Africa during the colonial era, examining the new social structures that emerged in cities to provide the necessary care, rest, and nourishment to enable people to show up at work each 5 See James Duminy's concise review of the historical literature in his 'Historical urban food governance in Africa: the case of Kenya, c. 1900-1950', in J. Battersby andV. Watson (eds.), Urban Food Systems Governance and Poverty in African Cities (London, 2018), 83. The classic text in the supply, marketing, and distribution framework is Janes Guyer's edited volume, Feeding the African City: Studies in Regional Social History, J. Guyer (ed.), (Bloomington, IN, 1987). See also C. Robertson, Trouble Showed the Way: Women, Men, and Trade in the Nairobi Area, 1890-1990(Bloomington, IN, 1997, and J. Rich, A Working Man is Worthy of his Meat: Food and Colonialism in the Gabon Estuary (Lincoln, NE, 2007). The scholarship we do have on cooking in African history principally examines food preparation in rural communities, such as E. C. Mandala, The End of Chidyerano: A History of Food and Everyday Life in Malawi, 1860-2004(Portsmouth, NH, 2004, and James C. McCann's survey of African food history, save for a brief passage on Ethiopian eateries, in Stirring the Pot: A History of African Cuisine (Athens, OH, 2009).  16, 52;and A. B. Trubeck, 'Kitchen work, 1920-present', in A. Bentley (ed.), A Cultural History of Food in the Modern Age (New York, 2016), 127-44. For rural African women cooking as part of domestic labor, see McCann, Stirring the Pot, and Devin Smart, 'Provisioning the posho: labor migration and working-class food systems on the early-colonial Kenyan coast ', International Labor and Working-Class History, 98 (2020), 173-92. 7 P. Scholliers, 'Workers' time for cooking and eating in nineteenth-and twentieth-century Western Europe ', Food and Foodways, 6:3-4 (1996), 244-6; I. Tinker (ed.), Street Food: Urban Form and Employment in Developing Countries (New York, 1997), 149-50; P. R. Duis, Challenging Chicago: Coping with Everyday Life, 1837-1920(Urbana, 1998Scholliers, 'Eating out', in M. Bruegel (ed.), A Cultural History of Food in the Age of Empire (New York, 2016), 108-9; A. Luissier, 'Eating out during the workday: consumption and working habits among urban laborers in France in the second half of the nineteenth century', in M. Jacobs and P. Scholliers (eds.), Eating Out in Europe: Picnics, Gourmet Dining and Snacks since the Late-Eighteenth Century (New York, 2003), 337-50; U. Thoms, 'Industrial canteens in Germany, 1850-1950and A. Carroll, Three Squares: The Invention of the American Meal (New York, 2013), 110-11. 8 Social reproduction theory examines the systems necessary to sustain working-class communities under capitalism on a daily and generational basis, keeping people physically, intellectually, and emotionally ready to work. B. Laslett and J. Brenner, 'Gender and social reproduction: historical perspectives', Annual Review of Sociology, 15 (1989), 382-3. day. However, we still know little about how these workers, once on the job, were able to labor from the early morning until the late afternoon. That is, the existing literature focuses on residential space as the site of social reproduction in colonial African cities. For example, the work of Frederick Cooper, Lisa Lindsay, and others shows the lengths to which the late-colonial state went to turn male migrant laborers into a more permanent and regularly employed urban proletariat, living in homes with wives whose unpaid domestic work reproduced their labor. 9 Other scholars like Luise White have demonstrated how sex workers sold a range of domestic services, which included cooking food for their clients. Their commodification of social reproduction, though, including the preparation of food, was tied to the residences where many sex workers ran their businesses. 10 Even research on 'market women', which has gone the furthest in explaining how the laboring classes in African cities acquired their daily food, mostly examines traders who sold produce, ingredients that had to then be prepared into meals at home. 11 All of this domestic and commodified labor in and around households was crucial to sustaining urban working classes in colonial Africa. However, by late morning in a town like Mombasa, hunger began to build for those doing manual labor in humid heat for the port, railway, and other urban employers. Many of these people had also commuted a long distance from their residential neighborhoods. Therefore, during working hours, they had to rely on nearby vendors who had spent their mornings chopping, boiling, and frying food so that a variety of dishes would be ready for the lunch-hour rush. Once satiated, customers could then return to their jobs for the rest of the day.
Those who ran these businesses were part of what is now called the 'informal economy'. Found throughout the world, the sector tends to expand when people are excluded from waged employment or if the quality and quantity of 'formal' work fails to keep pace with increasing numbers of job seekers. Facing these circumstances, people instead turn towards petty trade, which has low entry barriers because of its small capital start-up costs. The 'informalization' of the economy is also driven by growing classes of low-income consumers who seek out the cheap prices that these businesses offer because of their minimalist approach to overhead. 12 These factors that make the sector appealing to proprietors and customers alike are also the very features that historically have brought vendors into conflict with municipal administrations, around the world and in 9 F. Cooper, On the African Waterfront: Urban Disorder and the Transformation of Work in Colonial Mombasa (New Haven, 1987); F. Cooper, Decolonization and African Society: The Labor Question in British and French Africa (New York, 1996), L. Lindsay, Working with Gender: Wage Labor and Social Change in Southwestern Nigeria (Portsmouth, NH, 2003); and for similar labor policies starting slightly earlier, see G. Chauncey, Jr., 'The locus of reproduction: women's labour in the Zambian Copperbelt, 1927-1953', Journal of Southern African Studies, 7:2 (1981 Many sex workers did operate outside residential space, but they tended to only offer sex and not other services like food preparation. L. White, Comforts of Home: Prostitution in Colonial Nairobi (Chicago, 1990). See also M. Strobel, Muslim Women in Mombasa, 1890-1975(New Haven, 1979, 138-47; Chauncey, 'The locus of reproduction', 148; and J. Willis, Mombasa, the Swahili, and the Making of the Mijikenda (Oxford, 1993), 96.

11
For a classic study on food traders in Kenya, see C. Robertson, Trouble Showed the Way. In their monographs on Mombasa, both Strobel and Willis point to the key role of informal food vendors to social reproduction, including those preparing meals, but neither examines the role of street food within the spatial dynamics of the working day, nor do they push the analysis past a few passing comments, leaving the social history of street food largely unexamined: Strobel,Muslim Women,135,Willis,Mombasa,197. Strobel also published an influential essay on social reproduction in Mombasa, but it focuses primarily on the household, and examines the topic within the context of the precolonial city and slavery: M. Strobel, 'Slavery and reproductive labor in Mombasa', in C. Robertson , 1905-1970(New Haven, 1980 Kenya's port city. 13 In Mombasa, the central mechanism that kept overhead costs and therefore prices low at street-food eateries was their ability to avoid paying rent to occupy space, either to a commercial landlord or in the form of a fee for a hawker's license or a spot in a municipal market. 14 Instead, the city's vendors typically operated illicitly from push carts or in self-constructed shelters located on sidewalks, road reserves, and other open areas. This meant they were often unlicensed, asserted officially unrecognized claims to land, worked in structures that violated building codes, and ran eateries with limited access to sanitary infrastructures like running water and toilets. This was a business model that did not fit with the way urban planners imagined the future of Kenya's largest coastal town.
Nonetheless, municipal authorities largely allowed street-food eateries to expand and operate for decades, lacking the resources to implement their own laws and being anyway reluctant to disrupt the role these vendors played in social reproduction. However, by the late-colonial era, the state became focused on investing in and enforcing a new kind of urban order. As strikes and protests unfolded across late-colonial African cities, the British became concerned not only with 'labor stabilization', but also in asserting more control over the perceived 'disorder' of informal urban commerce. 15 During the final years of colonial rule in Mombasa, city authorities pursued campaigns to enforce licensing laws and fees, demolish roadside structures, and move vendors into stateconstructed facilities where they paid rents and could be more closely inspected. These policies may have brought about a certain kind of order, but by undermining the business model of street food they precipitated a crisis of social reproduction. This produced a contradiction. The underlying logic of late-colonial labor policy was to create a more stable and predictable workforce, but the full enforcement of the licensing, building, and sanitary laws of modernizing urban planners also made it harder for working people to access food while on the job. Indeed, it became quickly clear that taking away the affordable food that sustained people during the working day was not a particularly effective way to keep labor peace.
Positioned at the center of this contradiction, street-food vendors had considerable power in an unequal society, as the need to keep port, railway, and other workers fed in the middle of the day placed clear limits on how far campaigns to remove their roadside eateries could be pushed. 16 The Kenyan novelist Meja Mwangi recognized these dynamics in his fictionalized narrative of working-class life in 1970s Nairobi. In the novel, after city officials shut down street-food vendors near the construction site where his main character worked, they soon thereafter 'had to reopen. Someone realized that in order to keep them working, the workers had to be fed'. 17 In colonial Mombasa, municipal authorities may have wanted to fully implement their version of 'urban development', but both government and private employers needed working-class labor power even more.

Kitchen labor and urban capitalism
During the first decades of the twentieth century, street food was not a significant part of how Mombasa's workers accessed their daily meals. Until the 1930s, most labor migrants received a portion of their wages as a ration of raw foodstuffs, called posho. 18 Living on tight margins with limited cash earnings, they would have been reluctant to spend additional money to eat out when they could cook the food they already had. Moreover, Mombasa was still a relatively compact urban space during this period. People often worked comparatively close to their residence, which, whether they received rations or not, meant they could return to their domestic kitchen during the lunch hour to prepare, or be served, a midday meal. Businesses did exist that sold cooked food to Mombasa's poorer communities during these decades, but the consumer market for their services was limited. 19 These businesses included a vendor an official observed in 1919 who 'cooked Mahindi [maize], Mohogo [(sic) cassava], etc. … under the Mango trees' in the island neighborhood of Mwembe Tayari, or the three 'Swahili women' captured in an undated picture selling (seemingly fried) fish at small tables under a coconut tree. 20 Vending offered a living especially to coastal women who were excluded from most waged work other than domestic labor, and it provided quick and easy food to customers. 21 However, during this period, street food was not an essential component in the reproduction of the working day.
The number of street-food eateries, however, increased steadily as the city grew, especially from the 1930s onward. During that decade, food rationing largely ended, opening more cash income that working people could spend on prepared food. More important, though, was the fast-growing population of Mombasa. From 1930 to 1962, the number of Africans living in the city increased from 34,000 to 110,000, and most of them were in the laboring classes. 22 Population growth expanded the size of consumer markets for prepared food, and it transformed the city's geography in ways that made eateries more necessary. With tens of thousands of additional people living in Mombasa, the spatial scale of the port town extended outward. New African residential neighborhoods emerged on Mombasa Island further from the industrial district, and working-class suburbs also grew on the mainland, including Changamwe, Kisauni, and Likoni. 23 With longer distances between home and work, Kenya's port town became a city of commuters. Writing about the north mainland suburb of Kisauni in 1958, colonial sociologist G. M. Wilson described it as 'one of the major African dormitory areas of Mombasa' whose residents 'work mostly on the Island'. 24 These commutes pulled people away from their residential kitchens, expanding the market for cooked food in the middle of the day.
This process created demand. Structural transformations in Mombasa's political economy also increased the number of people looking to informal work like street-food vending as a way to make a living. The scarce labor markets of the first decades of colonial rule ended during the 1930s, and unemployment and underemployment became regular features of Mombasa's economy. By 1959, a government estimate reported that between 6,000-8,000 people were officially 18 Smart, 'Provisioning the posho'. 'unemployed'. 25 Additionally, such statistics referred mostly to men, and did not include the tens of thousands of women who were de facto excluded from most wage-earning work. Moreover, even though labor stabilization theory assumed that women would be supported by male 'breadwinner' wages, the reality was that even women in such households usually had to work, a situation Lindsay observed in colonial Nigeria too. 26 Indeed, a report on Mombasa from the late 1950s observed that 'a wife supplements the income of her husband by cooking for single men', along with other kinds of informal work, and generally 'it has been found that the majority of African women living [in the city] earn some type of income'. 27 Both men and women, responding to a difficult and gendered labor market, turned to the informal economy to make a living. Often, they opened street-food eateries.
The lunch hour was the core of the street-food business, but many vendors started the day cooking breakfast for people before work. Mombasa's residents could of course prepare their own morning meal at home, and often did. 28 In the city's many all-male households, this meant that men prepared their own food, a radical new departure compared to rural communities where women did nearly all the cooking. 29 However, women who lived in Mombasa's migrant households always cooked, which meant they and the men they lived with rarely ate breakfast outside of the house. 30 For many others, though, stopping by a vendor on the way to work offered relief from cooking and cleaning, as well as the opportunity to enjoy the novel flavors of a cosmopolitan urban cuisine. The principal breakfast food that migrants ate in their rural homes was a grain-based porridge, prepared by wives, mothers, grandmothers, aunts, sisters, and daughters each morning. Male migrants in the city had neither been taught how to prepare this dish nor did they have easy access to the sorghum and millet with which it was commonly prepared back home. Consequently, chaia caloric tea prepared with milk and sugaroften took the place of porridge in Mombasa's working-class diets, in households and especially on the streets. 31 Morning commuters could also complement their tea with a variety of dense, filling snacks from the coastal cuisine of the Indian Ocean. These included mahamri, square-shaped doughnuts made from a dough combining wheat flour with yeast, coconut milk, cardamom, and water, as well as kaimati, a similarly prepared small ball of fried dough saturated with syrup. For something less sweet, people purchased chapatti, the East African version of the South Asian wheat-flour flat bread, or simple prepackaged sliced bread. Both of these last two options took on more flavor when dipped into a sugary and milky chai. 32 The businesses that dominated the morning trade were gari ya chai, which were operated mostly by Arab and African men. Creating a commercial culture of fast breakfast cuisine, these vendors also 25 On unemployment in Mombasa and Kenya more widely, see: Cooper, On the African Waterfront, 46, 97, 210;CAC 753.14/v.19/1960 (4)  These conclusions rely to a certain extent on colonial-era evidence combined with oral interviews covering the independence period, a time for which interviews provide more detail on household dynamics. Interviews, 2014-15. In colonial Mombasa, women lived with migrant men as their wives, but also in more overtly transactional relationships called 'Swahili marriages' in which men paid them cash in exchange for their domestic labor. Willis, Mombasa, 108, 169. 31 Interviews, 2014-15. For integration of tea into rural diets among the Samburu, see J. Holtzman, 'In a cup of tea: commodities and history among Samburu pastoralists', American Ethnologist, 30: 1 (2003), 136-55. The classic study of highly caloric tea as working-class food is S. Mintz, Sweetness and Power: The Place of Sugar in Modern History (New York, 1986). 32 Conclusions about street-food cuisine during breakfast aggregated from UY files at KNA-Nairobi, as well as from interview with Nzalu Ngotho, Mombasa, 2014, and Deverell, 'A social investigation ', 62. contributed to the emerging urban architecture of economic informality. As their name in Swahili translates to 'tea cart', most of these vendors fittingly worked from hand-pulled wooden carts. Some were minimalist, a basic gari packed with tea, food items, and no further adornment. However, others constructed awnings onto their cart, offering shade from the sun or shelter from Mombasa's frequent rains, while another style attached the gari to a maduka (small grocery shop, in Swahili), in effect creating a hybrid formal and informal business. Rather than having to renovate and relicense maduka into fully authorized restaurants, shops could simply add a small cart outside and start selling prepared food. 33 Some tea vendors dispensed with the cart entirely, and ambitiously built more permanent illicit structures. In 1960, for example, Maganga Kashere operated from a small shelter 'made of canvas … supported on poles', while a different tea vendor, Machogo Ogola, had a 'tin shed' that he 'used as a restaurant'. 34 Another tea stand was in a '[t] emporary building' composed of 'one room divided by a cardboard partition'. 35 Buying breakfast from these gari ya chai was optional, as people woke up in their homes with access to a kitchen. However, when they started the working day, the temporal and spatial dynamics of employment in Mombasa greatly restricted the number of those who could reasonably return to their residence for a midday meal. 36 For the workers employed in the industrial district on the west side of the island, only residents of nearby Majengo could consider going home for lunch, assuming they wanted to when they could instead buy a cheap meal near work. However, for those living in more distant neighborhoods, this was not a choice. They relied on street food. 37 Indeed, by no later than 1929, petty entrepreneurs had established roadside eateries selling midday meals in the industrial district, with one official observing that a 'temporary market is held nearer the Port, about lunch-hour, where cooked foods are sold'. 38 During the 1950s, more than 100 vendors operated there, offering 'varying types of meals on the spot for the African workers'. Among these eateries, '[c]ompetition is keen, and prices have not … increased for many years. The portions are large, and the place is obviously popular with many of the men'. In 1958, a labor report underlined the role that street food played for mainland commuters who 'cannot get home and back in an hour [during their break]'. Consequently, Africans did 'not, as a rule, leave the Port area for a midday meal'. 39 The lunchtime menus of these eateries offered both coastal and upcountry cuisines. However, conspicuously missing was ugali, the maize meal porridge that had become the dominant staple for migrant workers. 40 In colonial Mombasa, ugali was a dish more commonly cooked at home, and was not widely available at roadside eateries. 41 Ugali suited the needs of men preparing a meal in their domestic kitchens, as milled maize was cheap, and its preparation simple: combine it with salt, water, and then boil, stirring as needed. Eating out, instead of repeating the regular fare, presented the opportunity for variety. The options from street-food vendors included githeri, a Kikuyu maize kernel and bean dishes that also resembled isio from the Kamba and nyonyo from the Luo. 42 From coastal cuisine, people could find rice, chapatti, grilled meat (mishkaki), fried fish, 33 KNA-Nairobi UY/12/698: 'Application for licenses ', 9 Jan. 1959, andminutes of HC, 6 Nov. 1959. and mbaazi, pigeon peas stewed in coconut milk. The East African versions of the Indian Ocean aromatic rice dishes pilau and biriani, however, do not seem to have been consistently sold by street-food vendors during the colonial era, only appearing regularly in the independence period. 43 Gari ya chai vendors, who served only hot drinks and snacks during the morning, also operated during the lucrative lunch hour, and expanded their menus. Along with caloric tea and fried dough, midday customers could also purchase grilled meats and boiled beans. 44 Other vendors offered a single item, such as grilled maize or roasted groundnuts. 45 These varieties of street-food businesses not only sold inexpensive prepared food in the industrial district, but throughout the larger city in 'innumerable wayside stalls'. 46 As the workday ended, commuters began the journey back to their residence. Like breakfast, dinner was a meal that could more easily be cooked and eaten at home after work, but vendors still offered services in the evening to those willing to pay. Like breakfast, these choices were shaped by a gendered division of household labor. Men eating at roadside eateries for dinner were more likely to live in households without women, while those who were married ate food their wives cooked at home, as observed by sociologist M. N. Deverell in 1947 noting that '[m]arried men would expect their wives to prepare at least one good meal per day'. 47  This was the case for full meals, but households of various kinds, including those with women, also used street food to ease the work of preparing dinner by purchasing precooked food. In an Indian Ocean town, this was most often fried fish, which were 'sold freely in the evening', and could be paired with ugali or rice prepared at home. 49 As in any political economy, kitchen labor in Mombasa drove social reproduction, converting raw ingredients into palatable meals that fed the community. Home cooks prepared most of the breakfast and dinner food in the city, as people woke up in their residence and then returned there at the end of the day. However, the spatial dynamics of urban capitalism required street-food vendors to take a leading role in providing lunch to the city's workforce. Consequently, these commercial cooks who served cheap food in convenient locations were essential to the reproduction of the working day in East Africa's most important port city. raise living standards, but also reestablish control over labor. 50 During this moment, municipal authorities also aimed to assert state power over how African communities used urban space. The scholarship on this period in Mombasa has focused primarily on housing, the way the state tried to create more legible forms of residential settlement for the city's working-class communities. 51 However, as in Dar es Salaam and Nairobi, municipal authorities in Kenya's port town also wanted to gain control over informal commerce on city streets, including roadside eateries. 52 Andrew Burton, writing about Dar es Salaam during this period, observed that 'the informal economy had no place in the orderly town colonial administrators endeavoured to shape'. 53 City officials, though, had to proceed cautiously. The projects they wanted to implement would transform the structures of social reproduction, whether by changing working-class housing or how people acquired daily necessities like food. Urban planners may have been displeased by the ubiquity of petty street trade, but informal commerce nonetheless fed a large proportion of the city's laboring classes. Indeed, the whole point of development, from a labor perspective, was to create stability, to bring people back to work and keep them there, to have cargo consistently moving on and off trains and ships. Whatever the modernizing ambitions of urban planners, Mombasa's wage-labor system had become quite reliant on informal eateries. 54 Removing the businesses that provided cheap, convenient, and filling food to people during long days at work threatened to cause only more disruptions to industrial relations in Kenya's port town.

Social reproduction and the limits of urban development
These dynamics explain why the municipality took rather measured steps to control street food for much of the colonial era. Mombasa's municipal board first firmly fixed its attention on urban eateries during the 1930s when it tried to establish a normative framework for 'restaurants' in its legal codes. 55 If not literally 'brick-and-mortar', the licensed restaurant was to be housed in structures made from solid and ostensibly permanent materials. The bylaws stated that '[t]he floor of any room in which food or drink are served' needed be made from 'cement or wood or some hard, impervious material', while kitchen walls should be 'covered with glazed tiles, cement or other hard material'. This sturdy structure also needed access to sanitary infrastructures, in particular 'latrine accommodation for the use of customers' and 'good and sufficient supply of wholesome water available on the premises'. 56 These building requirements effectively made the city's streetfood industry illegal, but many in the government also realized the impracticality of their enforcement in the city's existing economy. A 1936 municipal sub-committee reported that the laws, have not and are still not being strictly enforced on the poor class of natives, Arabs and Indians, as it is clear any attempt at strictly enforcing them would be not only a hardship on the applicants [food businesses], but also on the section of the public served by those applicants. 57 Municipal officials may have wanted to shape the development of eateries in Mombasa, but they also could not ignore the caloric needs of the workers whose labor ran the city's economy. 50 Cooper, On the African Waterfront and Decolonization and African Society. During the 1930s, therefore, the city government created laws defining what a 'restaurant' legally could be, but without regularly enforcing them. After the Second World War, while labor stabilization policies accelerated apace, government projects to reshape the informal economy remained more passive. There was much talk, as in 1947 when officials debated a ban on the sale of street food, and even the more radical idea that all informal trade 'be abolished completely'. 58 Actual policy, however, was less ambitious, like the experiment to have vendors work in the kitchens of existing liquor canteens, though done on far too small a scale to replace the services provided by street eateries. 59 No less prominent a figure than the Liwali Mbarak al-Hinawywho was the top non-European official on the Kenyan coastfelt that urban revitalization should not be pursued at the expense of assuring basic needs like the ability to access food, arguing that the municipal board 'should not only see to the cleanliness of the town but also to the welfare of its inhabitants'. 60 Take 'the case of the seller of roasted maize', al-Hinawy argued, 'to put that person out of business not only deprived that person of his livelihood but deprived a section of the community of a commodity it wished to purchase'. At the same meeting, Councilor Jetha also argued that 'the labouring classes cannot afford to go into restaurants', and so street vendors provided them 'roasted maize at a price within their means'. 61 Into the 1950s, such arguments in Mombasa's government were mostly persuasive, and prevented larger campaigns against informal vendors. 62 Localized efforts to clear street eateries off open land did occur, but even in these instances the approach was constrained by the need to keep workers fed. For example, in 1951 the municipal government expelled a large number of vendors by the Kilindini docks. Their customers responded quickly. Mombasa's dock workers, organized, militant, and with the ability to shut down the most important choke point in the region's commodity transport system, convinced the authorities to allow the eateries back and even to build a shed where they could be housed next to the Kilindini labor office. Despite this concession, management at the docks did want to maintain more control over how their land was being used. The shed was built for only 50 of the vendors, but soon thereafter 150 businesses had opened, which all 'provided … cheap and good food'. In response, the port removed 100 of them, trying to keep with the original design of the complex, even with reports that this led to a 'deterioration' in the quality of the food. 63 The inexpensive and convenient calories that vendors provided were crucial to maintaining the port's full working day, but officials were also keen to set limits on the scale of petty commerce on their land. Above all, though, dock authorities found it unthinkable to expel street-food vendors entirely from the area.
This model of accommodation between vendors and dock employers also inspired municipal officials to consider a privatized plan to formalize street food, one in which companies would build canteens inside their businesses where vendors could work. This was not such a radical idea for colonial officials, as the company canteen model was common in mid-century Britain. 64 In 1952, municipal authorities in Mombasa felt a similar system could emerge in urban Kenya too. The port now provided at least basic shelter for these eateries, and city officials also heard that nearby East African Breweries was considering installing their own canteen, 'a fine example' they thought. After the 100 vendors had been removed from the docks, city authorities were searching for a place to relocate them, and so they sent a circular to 'various companies who employ large [numbers of] African labour [ers] in the vicinity of the Kilindini Docks' hoping they would construct food canteens. These officials were soon disappointed. Some companies denied the request outright, while others vaguely 'referred the matter to their Principals in the United Kingdom'. 65 The municipality did not push the idea further, and it quickly faded from policy discussions. Indeed, only in 1961 did officials again query employers about whether they might build company canteens, and received a curt reply. Mombasa's Chamber of Commerce and Agriculture wrote back that 'the problem is not one for private enterprise'. 66 The city's employers were not interested in creating a new form of industrial paternalism that bound them to feeding their workers. Instead, they were content to pay them cash wages with which they would buy their own food.
The contradictory politics of street food For a large portion of the postwar era, then, neither the state nor private employers made a sustained effort to build an alternative infrastructure that could feed the city's workers. Consequently, there was no resolution to the contradiction between the desire for a city without the informal economy and the reality that street food was essential to social reproduction. Additional localized attempts to curtail street food were made, but most vendors operated with minimal interference for more than a decade after the Second World War. 67 This changed in 1957. That year, C. D. Rosenwald, Mombasa's Medical Officer of Health (MOH), pushed for rigorous enforcement of the city's restaurant by-laws. His specific reason was that vendors could potentially transmit typhoid within the town, but Rosenwald was likely also concerned about larger structural transformations. 68 Mombasa's African population had grown rapidly during the postwar era, and the city's racialized class system meant they were the very communities most in need of the services and employment opportunities that street food provided. In 1947, the African population of the city stood at 56,000 people. By 1958, it was nearly 90,000, which means there was likely a similar proportional expansion of the street-food industry. 69 Rosenwald, tasked with maintaining bylaws being ignored on an ever-expanding scale, decided to embark on an unprecedented campaign to rid the city of street-food vendors. 70 Councilors disagreed over this course of action, especially the few Africans in the city government, but their concerns were ignored. 71 In 1957, the council began removing vendors from the areas around Mwembe Tayari, 'clearing illicit food hawkers from the market'. 72 The operation then expanded to the rest of the city, and by August reference was made to the now 'rigid implementation of Municipal Bye-Laws which prohibits unlicensed food vendors from practicing within the Mombasa Municipality'. 73 The following month saw the continuing 'imposition' of laws that aimed to end 'food vending as hither [t]o practiced in the town'. 74 For decades, city officials had drafted laws aimed at restricting informal vendors in Mombasa, but refrained from enforcing them. Now, led by Rosenwald, the municipality was taking the radical step of removing the key services that these businesses provided to a majority of the population. Predictably, the expulsion of street-food vendors caused a crisis in social reproduction, as al-Hinawy and Jetha had anticipated almost a decade before.
As this became clear, officials scrambled to solve a crisis of their own making. In part, the government could not ignore the problems their policies created because the affected communities would not let them, using the new openings to state power created by late-colonial political liberalization. City residents lodged complaints with Ronald Ngala, Coast province's new member of the colony's Legislative Council. Taking up their case, Ngala wrote to Mombasa's government that '[m] any Africans have been to see me' who stressed 'the need for having some African food at Mwembe Tayari', an area particularly hard hit by the clearances. Ngala had visited the island neighborhood and saw that none of the formal restaurants in the area sold 'any African food, e.g., beans, sima [ugali], maize, porridge, e.t.c.', establishments that were at any rate 'too costly for the type of African wishing to have his meals at Mwembe Tayari'. 75 They also confidently pressured European officials. F. P. B. Derrick, Mombasa's district commissioner, both read Ngala's letter and 'received representations from a number of Africans' on the issue. Sensing the urgency of the situation, Derrick sent a circular to Mombasa's leading municipal authorities, including the town clerk, chairman of the Municipal Board, African Affairs Officer and, indeed, Rosenwald, the MOH. He noted the 'very considerable feeling there is among Africans in the town regarding the prohibition … on food vendors selling food in the streets and in open spaces'. Explicitly concerned about the implications for labor, he urged that 'we should go slowly in this matter as otherwise there may well be repercussions. There is little doubt that a very large proportion of the manual laborers employed in this town traditionally obtain their food from itinerant vendors', as they cannot afford restaurants 'where the food is much more expensive'. 76 Employers were concerned too. The branch manager of the Shell Company of East Africa, located in the northern Shimanzi quarter of the industrial district, requested that the MOH authorize a vendor to sell food to his workforce, someone who could 'prepare hot curry stew, japatties [(sic) chapatti], posho, cakes, etc'. 77 City officials had disrupted the system that delivered cheap prepared food to low-income workers without an alternative. In this moment of crisis, the political will materialized to invest in a more ambitious state-run infrastructure to provision cooked meals to Mombasa's working poor. These would be municipal food canteens where vendors could operate under the strict surveillance of the state rather than illicitly on the streets. However, this would take time, and in the interim an immediate compromise was needed to mitigate the effects that clearances were having on 'both food vendors and the masses of [the] African public'. 78 Facing pressure, councilors decided that banning the eateries that fed the city's poor residents without a replacement was untenable. By the end of 1957, there was a 'gentleman's agreement' that until sufficient food-vendor canteens existed there would be 'no serious enforcement' of the by-laws against illicit eateries. 79 That is, the government decided that the informal economy of street food would be made tacitly legal until an alternative basis for social reproduction had been established.
As plans for these canteens were debated, the municipality made clear that the first priority was to serve African wage earners. The construction of canteens was intrinsic to the labor question in Kenya's port city. Consequently, officials were especially concerned that a canteen be put in the industrial district. The attack on working-class eateries had also affected more residential neighborhoods like Mwembe Tayari and Majengo where plenty of people both lived and worked, but the immediate goal was to build a food-vendor canteen in the High Level area of the industrial quarter, home to the port and railway. 80 To construct a canteen, the city needed land in an area where the East African Railway and Harbors (EARH) owned most of it. Mombasa-based company officials understood the urgency of the situation, but the Nairobi office was less interested in losing control over valuable urban real estate. In their negotiations, Mombasa's government was blunt in stressing the disruptive potential of the situation. Officials urged that 'it is most essential in the interest of good labor relations … that a Food Vendors Canteen … be established in the Kilindini High Ground area with the least possible delay'. Many private firms in the area were also anxious that canteens be constructed 'to prevent labour troubles or difficulties', concerned enough that they donated £500 to build the canteen. 81 By the end of December, EARH's central authorities relented, and provided a space on Mozambique Road right by the railway tracks. 82 With at least one industrial site secured, the council turned towards expanding the network of city-run eateries throughout Mombasa. This included another industrial canteen in Shimanzi, just north of High Level, as well as others in the more residential areas of Mwembe Tayari and Majengo, and in the west mainland neighborhood of Changamwe. 83 However, the transition from policy approval to building and running these canteens was slow. As questions of land, funds, and fees were debated, vendors continued to feed the city's workers, protected by the 'gentleman's agreement' that their political organizing had forced upon the administration. 84 As months turned into more than a year, the limited capacity of the municipality to replace roadside eateries with city-run canteens became apparent. The issue was not simply the inability to construct sufficient facilities, though that was crucial, but also uneven interest among vendors in relocating themselves to these canteens. Some saw an advantage in taking their businesses off the street and into these spaces where they were less vulnerable to harassment from officials, but many did not. After a call for applications in January 1958, the municipality did not receive enough applicants to fill the proposed number of food-vendor stalls. 85 By September, interest had not increased, even as 'every possible publicity was made to try to get the right kind of people to apply'. Still, they had a 'very poor response', and were concerned the stalls might remain 'empty indefinitely'. 86 The reluctance of vendors to work in these canteens makes sense. Success as a street-food vendor came from keeping prices low and finding a location with a high concentration of customers. The people running roadside eateries avoided canteens because they required a license fee, adding to overhead, and they restricted mobility, tying the vendor to a place where they 80 were surrounded by competitors. They understood the economics of their businesses, and many stayed away. Impatient municipal authorities were beginning to understand this too.
In August 1958, the détente between city officials and illicit eateries came to an end, as the council reinitiated arrests and prosecutions. Vendors again pressured new African representatives in the government to help them. F. J. Khamisi, having recently assumed the Mombasa seat in the Legislative Council, wrote to the municipality and expressed his 'dismay and disgust' at these actions, and asked for an immediate suspension of the operation. However, Khamisi and his constituency's concerns were quickly dismissed in the name of sanitation, as 'the unlicensed food vendors were jeopardizing the health of Africans by not maintaining proper standards of cleanliness'. The health committee underlined that because Khamisi was advocating for illicit vendors rather than sanitation standards, he was proving himself a 'retrogressive rather than a progressive leader of his people'. 87 Sanitation, at least for these officials, superseded the immediate concerns of vendors and their customers. To be a leader in modernizing Mombasa, according to some British officials, was to understand this and enforce the laws.
However, a singular concern for sanitation fails to explain why arrests and prosecutions began again in 1958. Health standards were an equal part of the initial clearances in 1957, but African vendors and their representatives had successfully swayed both the district commissioner and then the council to pull back enforcement. What had changed was that officials now better understood that simply building canteens was not enough to convince vendors to relocate there. As David Harvey has written about such 'urban restructuring' in modern cities, '[v]iolence is required to build the new urban world on the wreckage of the old'. 88 In Mombasa, the creation of a new spatial and legal order meant that the old political economy of street food had to be destroyed. In the city, the existence of a large number of eateries operating with less overhead and the advantage to locate themselves in places most convenient for their customers siphoned off the consumers that otherwise would have had to frequent the municipality's newly opened canteens. Officials saw this, and so too did those vendors who signed on to work in these city-run institutions. This was made clear two months later when the council concluded that the further construction of municipal food stalls could only be justified if done alongside the mass prosecutions of those who operated illegally. Vendors working in the canteens also emphasized this when they told the council that 'their business was being seriously interfered with by the operations of illicit food vendors', and they could 'double their output if the demand were there'. The council agreed, and '[r]esolved that the licensed traders be protected by taking action to suppress illegal trading'. 89 By the end of the 1950s, health officials and a class of licensed eateries had agreed that the illicit economy had to be destroyed for the new one to take hold. 90 However, people still had to eat, and the municipal canteen system remained too small to adequately feed the tens of thousands of lowincome people who were in daily need of prepared food in Mombasa. By 1960, the council relented slightly in its campaign by making it legal to sell bread and mahamri, the Swahili doughnut, on the streets. 91 Customers, though, wanted to eat more than bread and doughnuts for lunch, and vendors were willing to provide them with additional options, even if it meant running their businesses illegally and exposing themselves to arrests and demolitions. These dynamics unfolded in places like Jubilee Square, a park next to the railway station both close to working-class customers and among the first places that arriving passengers to Mombasa saw. In 1962, the square had just been renovated, given a 'new layout together with a garden' that 'had considerably improved the general appearance of the area'. The municipality also then started a campaign of arresting and destroying the structures of the gari ya chai operating there who were 'ruining the appearance'. 92 By October, all the small structures had been demolished, but vendors nonetheless remained, continuing to earn their income by selling lunch to workers in the industrial district. 93 Municipal officials pushed forward the campaign against these vendors in Jubilee Square, while they also aimed on a more citywide scale to prosecute 'unlicensed food selling where possible'. 94 As independence approached, the people running Mombasa's street eateries increasingly found that a profession previously marked by small if steady profits was now defined also by insecurity, where any given day the capital they had invested could be destroyed.

Conclusion
In colonial Mombasa, the informal industry of street food was intrinsically tied to the formal world of wage labor. As the geographical scale of the city extended outward, vendors played a key role in reproducing working-class labor power by providing workers a lunchtime meal when they were far from their home kitchens. The steady expansion of street food and other informal businesses was also driven by a wage-labor market that could not easily absorb Mombasa's fast-growing population. Opening a street-food eatery, like selling fresh produce, used clothes, or engaging in other forms of petty trade, became an important way that people earned income as wage labor failed to keep pace with the large number of migrants moving to the port town. Even as structural incentives were pushing people into this kind of work, postwar urban authorities nonetheless decided that such forms of economic activity did not belong in a modern city. Municipal officials then gave this idea material force through their campaigns of fines, arrests, and demolitions.
However, the role of street-food vendors in the reproduction of the working day gave them power. Employers wanted people to perform hard manual labor from early in the morning until late in the afternoon, and access to an inexpensive, caloric meal was crucial to sustaining a content and efficient workforce, the principal goal of late-colonial labor policy. Thus, for much of the postwar era, the perceived disorder of informality was acceptable because it kept organized workers well fed and on the job. Moreover, people taking on informal trades like street-food vending eased pressure on the labor market, thus staving off an even worse unemployment crisis. This is the central contradiction of street food specifically, and informality more generally. By mid-century, these businesses had come to serve essential needs in the actually existing political economy, but their architectural style, tenuous legal status, and uneven access to sanitary infrastructures also rendered them incompatible with the vision of urban modernity held by many city administrators. As independence arrived in 1963, informal vendors and new African officials had to face the question of whether this contradiction could be resolved.
In the following postcolonial decades, the incentives that drove the street-food industry only intensified. Rapid suburbanization on the mainland cemented Mombasa's status as a commuter city, increasing the number of those needing a midday meal while away from home, while endemic unemployment meant that the rising urban population regularly turned towards the informal economy for income. The changing politics of decolonization did lead to reforms that aimed to mitigate the repression of street trade, but many independence officials still dreamed that Mombasa would become the kind of 'modern' city in which there was no place for economic informality. 95  enduring tension meant that the contradictory politics of street food continued to shape the lives of vendors after independence, like they did for Sampson Kereu, who recalled in a 2014 interview how economic security and precarity were bound together in his roadside business. Sitting in front of his eatery, Kereu talked about how the restaurant had provided him consistent income over the past four decades. He told me how customers formed long lines during the lunch hour to buy his food, particularly his specialty, the Kikuyu dish mataha (also called irio), a mash of bananas or potatoes mixed with beans and maize kernels. Among the more popular street eateries in Mombasa, Kereu explained the success of his business through the family life it allowed him to build. He was able to marry three wives, pay for his children's school fees, and to buy rural land back in central Kenya. It also provided a form of social security. In middle age, he began to lose his sight, which made working in the restaurant more difficult, but his employees kept the business running and earned him enough income to get by. 96 However, he also emphasized that operating a street eatery was 'risky', as harassment from the government was so regular it became routine. He remembered that often, sometimes even twice in the same day, he was arrested and his supplies were confiscated. After paying a fine, though, he could retrieve his pots and reopen the business until the next time the authorities came. 97 For street-food vendors, these quotidian incidents were punctuated by wider campaigns that sought to rid the city entirely of their trade through mass arrests and the demolition of thousands of roadside structures, as during a 1978 cholera epidemic or in the lead-up to the 2002 election. 98 Despite reforms to municipal laws, vendors remained exposed because most of them still occupied an ambiguous legal status, and even those who were fully licensed were granted only temporary access to the land where they ran their eateries. Such legal precarity facilitated the removal of vendors when government officials and other powerful constituencies decided it necessary, whether for public health, new construction projects, or to improve the electoral chances of particular political parties. Nonetheless, after these displacements, people consistently returned and rebuilt. Demolition crews smashed apart many thousands of structures made of wood, coral blocks, cardboard, and corrugated iron, but they did not change the underlying political economy that incentivized people to open these eateries and customers to frequent them. 99 In Mombasa, the contradictory politics of street food had made resilience a necessary condition to earn income as a vendor, as demolitions and other forms of harassment simply became part of running one of these businesses. Mohamed Amir Temu, who opened a roadside eatery in the 1980s, underlined this point. After a demolition, he explained, 'What did you do? You rebuilt again. It's destroyed one day, rebuild it the next morning'. 100 Acknowledgements. I want to thank Jim Brennan, Antoinette Burton, and Terri Barnes for their research mentorship over the years, on this and other projects. I am also grateful for the helpful comments provided by the editors and anonymous reviewers of The Journal of African History, which significantly improved the final article. Additionally, Onesmus Mutua and Mercy Nyambok were essential to the success of the interviews done for this project, as they organized and facilitated meetings with people who generously shared their experiences. Thank you. The bulk of the research for this article was funded by the Department of History and Graduate College at the University of Illinois at Urbana-Champaign, and subsequent trips to archives in Kenya and Britain were supported by the Department of History at West Virginia University. Time to write the original draft was provided by the Illinois Program for Research in the Humanities (now the Humanities Research Institute) while I was a Nicholson Fellow during the 2015-16 academic year.