Shea (Vitellaria paradoxa C.F. Gaertn.) – a Peripheral Empire Commodity in French West Africa, 1894–1960

HIGHLIGHTS Historical evidence indicates a widespread and centuries old exchange of shea kernels and shea butter by women in periodic local markets, and on a regional scale, with the densely populated West African littoral. Initial French interest in shea was as a potential substitute for gutta-percha (latex) to insulate submarine cables in the wake of the profligate, inefficient and unsustainable methods of extraction from Palaquium gutta and other tropical rainforest trees in Southeast Asia. Early and violent resistance to colonial rule during the Volta-Bani War was replaced by persistent migration into the (British) Protectorate of the Northern Territories of the Gold Coast Colony, as a form of protest by local Burkinabé communities to avoid the cumulative burdens of capitation tax, forced labour, military conscription, corporal punishment, and restrictive forest policies. Multiple initiatives to extract shea butter (mechanically and chemically), to protect shea parklands or to plant shea trees, as well as early industrialization efforts, including a French ‘colonial petroleum’ project, were not successful. Ultimately, the production and supply of shea kernels and shea butter remained central to servicing the needs of Burkinabe and West African consumers throughout the colonial period. SUMMARY Burkinabé women have traded shea kernels and shea butter in periodic local markets, and on a regional scale with the densely-populated West African littoral, for centuries. This paper traces the origins of French colonial efforts to develop shea as a commodity of empire from the 1890s to independence in 1960. Colonial efforts to incorporate Upper Volta, a French colonial backwater, into the world economy was drawn out, heterogenous, and messy. The colonial state assumed erroneously that little shea trade existed, and that producers would respond positively to market incentives. Yet, we suggest that French colonial policies failed due to a composite of factors including the limited investment in either the colony or shea as an oilseed crop, adaptation by women shea producers to the extraction of male labour and the trade opportunities created by new international borders, and the ‘blindness’ of colonial officials to the economic, social and cultural functions of periodic local markets used by women shea traders. The historical trajectory of the shea trade continues to have implications for current-day shea markets and their actors.


INTRODUCTION
Shea fruits, kernels and butter are Non-Timber Forest Products (NTFP) of the shea tree (Vitellaria paradoxa), the most widely occurring species in the agroforestry parklands of Burkina Faso (Boffa 2015). The butter extracted from their kernels serve as important ingredients in the diet of rural communities, and surpluses sold provide a critically important source of income to women. Shea is collected, processed and marketed by an estimated 18.4 million women across a 3.4 million km 2 belt across sub-Saharan Africa (Naughton, Lovett and Mihelcic 2015). For centuries, women managed shea trees in West Africa's agroforestry parklands as a food tree crop. Women collected and processed shea kernels to produce shea butter, the most widely-used oil throughout the region, and sold kernels and shea butter in periodic local markets often linked to larger regional trading routes. The customary management of these gendered landscapes (Carney and Elias 2006) and traditional uses of this vegetable fat were largely ignored during the colonial period as women were rarely, if ever, interlocuters with colonial administrators or scientists.
Current-day periodic local markets continue to perform a vital role in the economic and social life, particularly of rural women, throughout West Africa (Bromley 1971, Curtin 1971, Hodder 1971, Meillassoux 1971, Smith 1971, Hill and Smith 1972, and McKim 1972. In Burkina Faso, a complex hierarchy links periodic three-day village markets to larger foci of trade in places such as Leo, Tenkodogo, Bittou and Koudougou and Bobo Dioulasso (Rousseau, Gautier, and Wardell 2015). In turn, these are linked to centres of consumption in neighbouring Ghana, Côte d'Ivoire, Benin and Togo, and other parts of the West African littoral. The aggregation of the periodic market system facilitates trade in surpluses of agricultural food and cash crops such as millet, yams, onions, tomatoes, livestock, shea nuts and butter (Hill 1970, Clark 1994, House-Midambe and Ekechi 1995, Chalfin 2004a. Given the promise shea holds for strengthening rural women's incomes as well as national economies in West Africa, several actors have attempted to radically transform shea value chains originating in West Africa by introducing new technologies, building capacities of women producer groups and improving access to new markets. These include government agencies, development and Non-Governmental Organizations, Trans-National Organizations as buyers of shea kernels and third-party certification bodies. Their efforts focused on relatively recent events and actions, detached from the broader historical context in which they are embedded. This, "at a time when short-term horizons constrict the views of most of our institutions, whether governments, non-governmental organizations (NGOs), corporations, or, increasingly universities" (Armitage and Guldi 2015: 221). History matters. It matters not just because we can learn from the past, but because the present and the future are connected to the past by the continuity of a society's institutions (North 1990, see also Huillery 2009).
Hence, this paper takes an historical perspective to contextualize current day shea value chains, focusing on the period after the establishment of the largest possession of imperial France -Afrique Occidentale Française (AOF), a federal constellation of colonial states -in 1895. The paper aims to show that women were instrumental in sustaining rural livelihoods by using local periodic markets to trade shea kernels and butter on both sides of the new borders separating Upper Volta from its neighbouring territories. After presenting the analytical framework and methods adopted, the paper initially presents an overview of pre-colonial patterns of shea trading drawing on archaeobotanical sources, and the records of travellers and explorers between the 14 th and 19 th centuries. This is followed by a review of the early interest in shea by the French administration, prior to the creation of Haute Volta as a colonial state in 1919, and the modest growth in the shea trade during the colonial era. A discussion of the reasons for the failure of colonial efforts to develop shea supply chains precedes the concluding remarks.
The paper should be read in conjunction with another paper 'Shea -the emergence of Global Production Networks, 1960-2021' (Wardell et al. 2021), which traces the postindependence governance of shea supply chains in Burkina Faso. The two papers help to explain the gendered nature of the contemporary shea economy in Burkina Faso by understanding the patterns of pre-colonial and colonial trade in the commodity. Women shea nut producers and shea nut processing in Upper Volta lived in a "world that didn't count" (McMichael 2004: 7), and remained in the shadow of cocoa growers in Côte d'Ivoire and Ghana until the post-2000 boom in the global shea trade. Women remained, nevertheless, firmly in control of the local and regional shea trade. Ironically, the key factor which influenced the placement of the border separating British and French territories was the protection of metropolitan trading interests.

ANALYTICAL FRAMEWORK
Landscape histories have important implications in understanding forest ecology and social and political relationships within current agroforestry landscapes, and in identifying coping strategies and adaptation to environmental stress. Historical perspectives can also increase our knowledge of the dynamics of tropical dry forest landscapes and provide a frame of reference to assess contemporary patterns and processes. Historical records in Africa, however, are often fragmentary. Even where longer historical time series can be assembled, the selection of appropriate reference conditions may be complicated by our limited knowledge of the past influence of humans, and by non-equilibrium dynamics. These complications do not lessen, however, the value of history; rather they underscore the need for multiple, comparative histories from many locations for evaluating both cultural and natural causes of variability. The reconstruction of landscape histories which recognize hierarchical scales of analysis in both time and space can highlight the complexity of specific local geographical and historical settings, and provide a basis to redefine baseline ecological conditions, to reinterpret the impact of population growth or, as one scholar has suggested, to "…systematically build in perspectives from political economy as well as ecology…" (Beinart 1996).
Environmental history -as a conceptual and interdisciplinary arena (Dovers 2000) in which 'humans are restored as part of nature' (Myllyntaus 2001) -provides a useful framework to cross conventional boundaries between the natural and social sciences, and to give more sympathetic attention to the complex mosaic of different land and resource users in African agricultural landscapes (Andersen and Grove 1987, Batterbury and Bebbington 1998, McCann 1999, Beinart 2000, Hays 2001, and Beinart and McGregor 2003. Some researchers have described the lasting influence of colonial regimes in shaping the contemporary institutional arrangements for the conservation of forests in West Africa (Fairhead and Leach 1995, Buttoud 1997, Ribot 1995, 1999aand 1999b, Firmin-Sellers 2000, Becker 2001, and Saul, Ouadba and Bognounou 2003. Other scholars have used historical data to question forest conservation 'orthodoxies' (Leach and Mearns 1996, and Fairhead and Leach 1996 which failed to recognise the important roles played by local farmers and the extent to which they were influential in enriching agricultural and woodland landscapes (see also Richards 1985, andBaker 2000).
Complex processes of differentiation and primitive accumulation have characterised West African production systems for centuries (Hopkins 1973, Wilks 1982, and Lovejoy 1985. However, processes of commoditization in African farming systems have often been ignored as an explanatory variable to help explain environmental change (Bernstein and Woodhouse 2001). Furthermore, recent approaches to global commodity chains (GCCs) and global value chains (GVCs) moved away from earlier long-term world-historical perspectives on commodification Wallerstein 1977, Hopkins andWallerstein 1986, see also Landsteiner and Langthaler 2021). Emphasis has frequently been given to quality regulation, restructuring processes, standard-setting, upgrading and the institutional 'rules of the game' Schmitz 2000, Raikes andGibbon 2000), and more short-term and industry-and firm-centred analyses (Bair 2008, andGrewe 2019). 1 In contrast, scholarship on Global Production Networks (GPN) emphasizes the multi-scalar dynamics of globalization, particularly in terms of the embeddedness of global networks in national, regional, and local contexts, encompassing both state and non-governmental actors (Coe, Dicken, and Hess 2008). Commodities have led to the consolidation of a globalized economy and societies but this has often been forged out of distinctive local experiences of cultivation and production, and regional circuits of trade (Curry-Machado 2013, andHazareesingh andCurry-Machado 2009: 1-5).
This paper, by focusing on a francophone colonial backwater, is a corrective to the focus in much environmental history on both anglophone 'settler colonies' (Ofcansky 1984, Beinart and Coates 1995, Griffiths and Robin 1997, and Brown 2003, or on the timber producing and Guinea-Savanna regions of West Africa (Gillis 1988, Wilks 1993, Parren 1994,

A BRIEF HISTORY OF TRADE IN WEST AFRICA
Historical patterns of local, regional, and international trade in West Africa pre-date the arrival of the Portuguese on the West African coast in the 15th century (Brooks 1993, Wilks 1982, and Hopkins 1973. The ebbs and flows of the dominant commodities exchanged during the last five hundred years have been determined by a complex web of social, cultural, religious (notably the expansion of Islam and later Christianity), economic and ecological factors across the region (Fage 1969, Daaku 1970, and Webb 1995. These, in turn, were influenced by the larger framework of the world economy, progressive improvements in transport and communications technology, and the influences of fluctuating demand and commodity prices. It was, for example, the opening up of new maritime markets in West Africa after the Portuguese trading monopoly was overthrown in 1642 which resulted in the progressive decline in trans-Saharan trade in gold by the late 17th century (Wilks 1982 andBrooks 1993). European penetration of the West African hinterland did not begin in earnest until after 1900 (Brooks 1970 andLynn 1998).

Linkages between local, regional, and long-distance trade
A multitude of short distance commercial networks and longer distance trading relationships have continually forged linkages between local, regional, and global trade (Newbury 1972). Furthermore, long-distance trade was possible long before the advent of colonialism as common monetary units such as gold standards and cowries existed, and trade relays involved multiple exchanges and modes of transportcamels, donkeys, and human porterage -in different ecological zones (Webb 1995). Elaborate decentralized networks of commercial alliances were maintained as trade routes changed, and new loci of commercial centres emerged such as Zuarungu close to present-day Bolgatanga in northern Ghana. Some scholars have highlighted the persistence of local and regional trade in, for example, salt (Sutton 1981), kola nuts (Lovejoy 1980 andAbaka 2005), and shea nuts (Chalfin 2004a and2004b).
European trade was largely confined to the West African littoral until the early 20th century. It was nurtured, initially, by African middlemen and European firms who operated as personalized oligarchic semi-monopolies. A significant expansion in the export of bulk commodities from, and bulk imports to West Africa occurred during the 19th century (Davies 1976 andArhin 1980). This period -often characterized as the era of 'legitimate trade' after the abolition of slave trading (Mcsheffrey 1983) -was dominated by the growth of trade in palm oil and palm kernels in exchange for imported cloth, guns, gunpowder, iron, salt and rum (Brooks 1970, Chamberlain 1979, and Lynn 1998. The transition to bulk Ibo andLeonard 1997, andKotey et al 1998). Furthermore, little scholarly work has been published on the history of shea as a colonial commodity although excellent work has been published on the contemporary governance of shea value chains (for Burkina Faso, see Carney 2004, Rousseau et al 2005, Elias and Arora Jonsson 2017, for northern Ghana, see Chalfin 2004a andGilli et al 2020). By adopting a Global Production Network approach to the evolution of the shea trade in Burkina Faso, the two papers emphasize the emergence of global trade which has been forged out of distinctive local experiences of production and transformation and local and regional circuits of trade controlled by women.

METHODS
The paper draws on a literature review and archival research primarily conducted at the: i. Archives Nationales, Ouagadougou (Burkina Faso); ii. Centre National des Archives Outre-Mer, Aix-en-Provence (France); iii. Institute of Forestry, Agricultural and Environmental Engineering in France, Nancy (France); and iv. Archives Nationales in Abidjan (Cote d'Ivoire). In addition, due to the large-scale exodus of Burkinabé people into (now) northern Ghana throughout the period after the Volta-Bani War (1914)(1915)(1916) and the re-creation of Haute Volta in 1947, additional research was conducted in some anglophone archives, notably the Public Records and Archives Department (PRAAD), Tamale (Ghana); Rhodes House, University of Oxford (United Kingdom) and the University of Edinburgh Special Collections (Scotland).
In general, better archival records exist for commodities such as cocoa and palm oil given the larger quantities and, hence greater values traded with European metropoles (Kotey et al. 1998, Lynn 1998, and Acquaah 1999. 2 We suggest that the comparative lack of historical data regarding the shea trade also reflects the complexities of unrecorded international trade with shea products moving south and north across international borders after colonial boundaries were first defined in the early twentieth century (Ellis and MacGaffey 1996, Fold and Reenberg 1999, Chalfin 2001, and Nugent and Asiwaju 1996. The use of multiple archival sources from French and British colonies facilitated understanding of how shea women were able to spread labour demand for processing and selling shea products throughout the year, and to use the opportunities provided by an elaborate network of periodic markets and trade relays to the south and north of the (new) international boundaries between Upper Volta and neighbouring countries such as the Northern Territories of the Gold Coast Colony, Côte d'Ivoire and Benin. The creation of a subversive borderland economy is explored further in the Discussion section of the paper.
2000-2500 years ago as patterns of climatic aridity were moving southward (cited in Lovett 2015: 134). Additional archaeological evidence suggests that the processing of shea in western Burkina Faso has occurred for more than two thousand years (Gallagher, Dueppen and Walsh 2016).
The earliest recorded references to shea date from the fourteenth century as trading relations between the Mali empire and Egypt grew ( Al-'Umarî 1927, andLevtzion andHopkins 2000). Muslim travellers travelled along trade routes that crossed the Sahara, and included Al-'Umarî born in Damascus. His travels in ca. 1337-38 led to the encyclopaedic 'Pathways of Vision in the Realms of the Metropolises'. 4 Subsequently, Ibn Battuta born in Tangier, undertook a 24-year long world-wide trip through most of the Muslim world. His last trip across the Sahara to the Western Sudan (ca. 1352-1353) led to the Rihla (Journey) a co-production with a scribe which also described uses of shea oil. 5 The arrival of European explorers to West Africa's interior from the end of the eighteenth century brought additional attention to shea. Mungo Park's search for the source of the Niger River took him along overland trade routes from the Gambia River into Guinea and Mali. The Scottish explorer noted the importance of the edible shea butter trade to Mandespeaking populations, and the product's multiple uses in his book, Travels in the Interior Districts of Africa, published in 1799. 6 By the end of the nineteenth-century, French explorers, French envoys of the British Crown, and colonial military "aménageurs" had all contributed to the considerable detail on shea that would later help in shaping its potential role in international trade. Colonial "aménageurs" were frequently hired to explore and administer new colonial territories prior to the establishment of civil administrations (Puyo 2001: 484). This was equally true for the Gouvernement général de l'Afrique-Occidentale française (GGAOF), where the French had already established treaties with chiefs along the Senegambia coast since the 19 th century. The expedition of René Caillié in the region from 1824-28 provides historical detail on the value of shea among the Mandingo (Caillié 1965(Caillié (1830. 7 Joseph Dupuis, a French envoy of the British King George the Fourth, described some of the tricks of the shea trade in his Journal of a Residence in Ashantee (Dupuis 1824). exports was facilitated by falling prices for bulk steamship transport, the advent of the telegram, and growing demand for raw materials amongst the new industrial economies of Europe. 3 The period did, however, usher in some changes notably in terms of the organization of European trading firms by separating the functions of merchants and shippers and progressively shifting to more corporate forms of ownership and management. Other efforts aimed to bypass coastal agents to reach African markets in the hinterland, and the increasing 'democratization' of trade as smaller traders increasingly entered markets. It was also an era characterized by classic 'boom and bust' cycles and periodic failures associated with trade in palm oil in the 1860s (Milbourn 1970, andLynn 1998), and mining in the 1890s (Silver 1981). During the latter part of the 19th century, trade in wild rubber (Arhin 1972) and West African mahoganies (Kotey et al. 1998) started to expand.

Pre-colonial patterns of shea trading
Shea provided the primary vegetable fat over a vast region of semi-arid Africa, in which the African oil palm Elaeis guineensis, used for similar purposes in wetter environments, does not grow (Lewicki 1974). Shea butter was traded as far south as the Volta River in Ghana for products from the coast (salt and fish) and forest (kola nut) (Ehret 2002: 321-22, andSutton 1981). Mossi, Juula, and Hausa caravans carried shea butter over long-distance trade routes prior to the Atlantic slave trade (Lewicki 1974: 106, Park 1983(1799, and Barth 1968 (1857)). First-hand accounts from slave traders operating along coastal Gambia and Guinea-Bissau in the early sixteenth century also confirmed the shea trade from the interior to the Atlantic coast, where the tree did not exist (Terpend 1982). The demand for shea butter increased with the slave trade as it provided a moisturizer for Africans awaiting sale to Europeans (Cowley 1928).
The earliest evidence of shea trees being utilized is from an archaeological site in northern Burkina Faso dated to ca. 1000 AD (Neumann, Kahlheber, andUebe r 1998, andKahlheber 1999). Hohn and Neumann (2012) suggest, furthermore, that agroforestry parklands intercropped with millet began to be cultivated in the north of the country about Louis Hecquard (1855: 373-74) commented favourably on the taste of the shea fruit, and the ability of the oil to last years without turning rancid. David Asante's observations in 1877 of items traded by the Hausa and Mossi caravans in Salaga included "a sort of vegetable oil which is firm even during hot spells" (Johnson 1966, cited in Duperray 1999. Joseph-Simon Gallieni's (1885: 440) exploratory mission to the upper Niger River in 1879-81 provided an early observation that women and children collected shea fruits. It was Paul Soleillet, however, who was the first person to note that shea butter (French: karité) was prepared exclusively by African women, a detail that was missing from all previous descriptions of karité (Soleillet 1887, cited in Carney andElias 2006: 242).
In the same year, Sir Alfred Moloney, Governor of Lagos Colony between 1886-1891, published his Sketch of the Forestry of West Africa with Particular Reference to Its Present Principal Commercial Products. He noted that shea "constitutes a main article of inland commerce, and is employed by the natives for anointing their bodies, for lighting, and for food. This fat is imported from Sierre Leone to the extent of 300 to 500 tons annually for use in the manufacture of hard soaps, chiefly in combination with other oils. A substance somewhat resembling gutta-percha is found in shea butter and is called Gutta-Shea; no application, however, has been found for it" (Moloney 1887: 379, see also Omosini 1975). This was in contrast to early French interest in shea as a new source of latex.

EARLY FRENCH INTEREST IN SHEA -THE ORIGINS OF THE OVERSEAS SHEA VALUE CHAIN
The largest possession of imperial France -Afrique Occidentale Francaise (AOF hereinafter, French West Africa) -emerged as a federal constellation of colonial states in the late 19 th and early 20 th centuries (Guernier 1949, Coquery-Vidrovitch andMoniot 1993) (Figure 1). Forts and trading posts established in the 17 th century on the West African coast provided the springboard for territorial conquests in Sénégal, and the later advances into the Western Sudan and southward to the Gulf of Guinea (Webb 1995).
The period after 1895 witnessed significant social, economic, political and environmental changes throughout the region as African communities were confronted with increasing demands for labour, commodities, and African territory. The scrambles for territory and commodities focused initially on the delineation of boundaries around the so-called 'Neutral Zone' (Figure 2). This encompassed efforts by the French, German and British to negotiate Treaties of Friendship and Trade with customary chiefs. George Ekem Ferguson, a Fanti official of the British Crown, played a key role in securing treaties of friendship and trade with many traditional rulers in northern Ghana and the neighbouring French territories of Haut Sénégal and Niger, Côte d'Ivoire and Togoland (Metcalfe 1964: 457, Thomas 1972, Arhin 1974, and Brukum 1997. These included a treaty negotiated in July 1894 with the Mogho Naaba, the leader of the dominant Mossi ethnolinguistic group. 8 These treaties aimed to grant the British free FIGURE 1 French West Africa, 1895-1960(Source: Cohen 1971 8 Binger had earlier tried unsuccessfully to negotiate a treaty with the Mogho Naaba for the French protectorate (Binger, 1892). access to the areas concerned, and the right to conduct trade. 9 In practice, boundary disputes concerning the 11 th parallel between the French and British administrations continued for many years and culminated in a joint British French mission (O'Kinaely-Verlaque) in 1904. This resulted in the delimitation of the boundary and a memorandum which recognized the user rights to lands belonging to all 34 villages straddling the boundary (Duperray 1999: 137-142). The communities were, in addition, granted the possibility to stay or to move to the other side of the boundary within a year. 10 It is not known if these rights included rights to use shea trees.
Early French tropical agricultural research was characterized by the establishment of several small botanical gardens after 1880, which accompanied the (military) conquest and early occupation of France's West African possessions. 11 This led to the founding of the Jardin colonial in Nogent in 1899, initially as an annex of the Natural History Museum in Paris. After 1900, new arrivals from the French school of agronomy   FIGURE 2 The 'Neutral Zone' between British, German and French territorial interests, 1898 9 The establishment of these 'free-trade' agreements led to an escalation in the military presence in the region in the late 1890s. A British military detachment led by Major H.P. Northcott occupied Gambaga where he was subsequently appointed Commissioner and Commandant in the Northern Territories of the Gold Coast. The defeat of the Ashanti in 1900 also heralded the beginning of a new trading era in the Gold Coast hinterland which soon set great store by the potential of the shea nut. 10 Archives nationales de Côte d 'Ivoire, 2 E 93 III, 8, 76. Memorandum franco-anglais 1904. 11 Similarly, Alfred Moloney, Governor of Lagos Colony, 1886-1891 set up a botanical station in 1887. This experiment was later accepted by the British Colonial Office as an important cornerstone of economic policy for the whole of British West Africa in the 1890s when the demand for tropical raw materials increased sharply. He also drew early attention to deforestation in both Lagos and the Gold Coast Colony (Omosini, 1975: 665-669). in Kati (present-day Mali) to continue research on shea. Jean Vuillet's first report -Etude du Karité considéré comme producteur de gutta -was addressed to the Delegate of the Governor-General (Vuillet 1901). He published a much more detailed monograph -Le Karité et ses Produits -for the GGAOF in 1911 (Vuillet 1911), which recognized the need to provide additional protection to shea trees and to support the establishment of shea plantations. 13 The extension of political control inevitably raised the issue of ownership, management and access to land and forests. Four decrees were adopted on 20 July 1900 defining the forestry, land tenure, State lands and public property régimes in Sénégal and its dependencies. By late 1900 the Governor had powers "….to set aside such parts of the forest as are already being exploited whose preservation he may deem useful, until further notice." In 1904, following the French Civil Code, the colonial administration decreed that all "lands that are vacant and without master, in the colonies and territories of French West Africa, belong to the state".
Early Colonial state restrictions on the use of both woodfuels (primarily for the railways) and Non Timber Forest Products (NTFPs) including shea were adopted in 1906, 1912, 1914 and 1918. 14 A GGAOF Forestry Code was adopted in 1935. Legal instruments of political-administrative control in Haute Volta created a history of fear (Ribot 1999a, see also Ribot 1995 andRibot 1999b). Shea nuts were also collected for the colonial administration using forced labour (corvée); women bore the brunt of this task (Massa and Madiega 1995, see also Madiega and Nao 2003). 15 Forced labour continued until 1946 with the more populated cercles in Haute Volta affected the most.
French interest in the export of shea products was limited at this stage given unfavourable customs import duties on oilseeds and competition from French farming interests which were advocating expansion of the domestic production of oilseed rape (Pehaut 1974: 445). The British, Germans, Dutch and Belgians were, in contrast, more interested in shea as a commodity for the margarine industry. 16 In 1907, a British company expressed interest to the Governor General of led to experimental trials on larger areas to compare varieties, as well as soil preparation and fertilization methods (Bonneuil and Kleiche 1993). The first agricultural station was opened in Banfora in 1904 with research focusing on rubber produced from the liane Landolphia heudelotii (Tourte 2019a: 153).
Initial French in terest in shea was as a substitute source of latex (gutta-percha) used as an insulator of submarine cables following the "profligate, inefficient and unsustainable methods of extraction" of Palaquium gutta and other tropical rainforest trees in South-East Asia (Tully 2009: 559, see also Anon 1892, Kennedy 1971, andPotter 2005). Heckel, who later became the Director of the Museum and Colonial Institute in Marseille, published the first (known) article entitled 'Sur un arbre producteur de gutta et de corps gras' (Heckel 1885 and1897). This led to further studies of shea by Dr. Rancon in 1891 and later Coppin at a botanical garden in present-day Mali. The great French botanist Auguste Chevalier undertook eight missions in AOF between 1898-1912(Chevalier 1902, and 1912 which led, initially, to the publication of his seminal tome Végétaux utiles de l'Afrique Tropicale Française (Chevalier 1905, and1917, see also Planchon 1888) 12 and later, the first forestry and pastoral map of AOF (Chevalier 1912). Samples of shea gutta were first sent to the French Chambers of Commerce in Le Havre and Marseille in 1897. A year later a veterinarian Cazalbou, forwarded cases of shea butter to La Maison Poncelet in Marseille. Growing interest in shea as a potential new source of either gutta-percha or as a raw material for soap manufacture, saw several studies published by Henri Jumelle and others in the late 1890s (Vuillet 1901: 8-11, see also Dastre 1901).
There are no archival records of early British interest in shea as a source of latex. In 1895 a first shipment of 23.7 tonnes of shea butter was exported by the British from Lagos to London and Liverpool (Pehaut 1974: 444). Regular shipments started in 1905 and once the railway linking Lagos and Kano was completed in 1912, the trade increased from ca. 3.000 tonnes in 1905 to more than 9.600 tonnes in 1913 primarily for use in soap and margarine manufacturing.
In 1900, Ponty, then Délégué du Gouverneur-Général du Haut Sénégal et Moyen-Niger, requested the agronomy station Haut-Sénégal-Niger, in buying 100.000 tonnes per year of shea nuts on the basis of a five-year renewable contract. 17 By 1910, the Haut-Sénégal-Niger colony had started to export shelled kernels rather than whole nuts. This improvement, by reducing the volume of exports per unit mass of kernels, led to an increase in kernel exports, from 25 tons in 1910 to 243 tons in 1911 (Perrot 1915). Margarine manufacturers were the first outlet for shea exports (Pehaut 1974) but shea was also used in soap and candle manufacturing as well as in pharmaceutical products and as an industrial lubricant. In the 1920s, irregularities in consignments, inconsistent product quality and transport constraints from the AOF's landlocked countries hindered expansion of the export market. No shea plantations were successfully established during the colonial period in either French or British colonies (for northern Ghana see Wardell and Fold 2013).

Mode st growth in the shea trade in the early colonial period
The interest of colonial trading firms and early industrial investments led to modest growth in the shea export trade in the early colonial period in Upper Volta, which served primarily as a labour reserve for French West Africa. Trade in shea by women continued in local periodic markets in Upper Volta and across borders in neighbouring territories.
Upper Volta as a 'nation state' had a chequered history after its creation in 1919 (Izard, Bonnefond and d'Huart 1967, Engberg-Pedersen 1993, McFarland and Rupley 1998, Englebert 1996, Duperray 1999, and Chafer 2002, following advice from Maurice Delafosse 18 to separate the vast Haut Sénégal-Niger territory (Royer 2003: 44). The country was suppressed in 1932 and the territory was redistributed between Haute-Côte d'Ivoire, Niger and Soudan Français (Mali) (Kobi 2019). 19 In the same year, Colonial Inspector Bernard Sol noted that "….one part of the Agricultural Service appeared to have been completely neglected: we are referring to the forestry sector". 20 The first Code Forestier for GGAOF was adopted during this period in 1935. Haute Volta was reconstituted in 1947 and gained full independence in 1960. The country adopted its new name, Burkina Faso, after Thomas Sankara came to power in 1983 (Rothschild and Gyimah Boadi 1989). After World War II, the colonial state's interests shifted towards groundnuts and cotton as the priority crops to underpin the colony's economic development. Low investment, limited transport infrastructure, changing metropolitan demands for oilseeds, and the ability of local producers to continue to use their own local and regional marketing networks also contributed to the decline of the shea nut export trade in the late colonial period. In essence, the production and supply of shea products remained central to servicing the needs of Burkinabe and West African consumers throughout the colonial period.
The Burkinabe population was, inevitably, affected by the establishment of colonial states, and institutions such as Forestry Departments, as well as efforts to integrate local production systems into the global economy. These forces interacted continuously, however, with long-established patterns of customary land and resource use, migration, social change, and internal trade (Gregory 1974, and Cordell, Gregory and Piche 1996. Early and violent resistance to colonial rule was substituted by persistent migration from the region as a form of protest by local communities to avoid the cumulative burdens of capitation tax, forced labour, military conscription, corporal punishment, and forest policies (Asiwaju 1976, Saul andRoyer 2001, see also Dresch 1945 andRouch 1956).
Lieutenant Governor Hesling considered the territory of Haute Volta to be devoid of real forests shortly before its suppression in 1932 21 before the belated introduction of the 'empire forestry mix' 22 (Barton 2002) in Haute Volta. This led 17 Lettre de la maison Eduard de Val and Co., 3, East India Avenue, London EC, 30 January 1907 a M. le Gouverneur du Haut-Senegal-Niger cited in Pehaut, 1974: 444. 18 Maurice Delafosse headed the Bureau of Indigenous Affairs in Dakar and was renowned for his cultural approach to African history. 19 Prior to the recognition of Haute Volta, it was known as the 2 nd Military Territory of the Soudan (1898-1902), then formed part of the colony of Sénégambie-Niger (1902-1904, and subsequently, part of the colony of Haut- Sénégal-Niger (1904-1919 The 'empire forestry mix' was first developed in colonial India and Burma prior to being exported to other anglophone colonies around the world. It comprised three key elements viz. the appropriation of large areas of land to establish national networks of protected areas (National Parks, Forest Reserves, Game Controlled Areas etc); the regulation of timber and Non Timber Forest Products to generate revenues to help pay for colonial administration; and the introduction of 'scientific forestry' to manage forest resources by a state administration (Barton, 2002, see also Grove, 1995 andWardell, 2020).
to the appropriation of large areas of land by the colonial state: seventy protected areas were gazetted in accordance with the 1935 federal Code Forestier before the country gained independence in 1960 (Ribot 1995and 1999a, Wardell and Reenberg 2006. 23 As renowned Burkinabe researcher Edouard Bounkoungou noted in 2005: "Dry forests in arid and semi-arid regions were viewed with disdain as scrublands of slow-growing twisted bushes, which benefited from little attention as they represented no value to the commercial timber trade" (Bounkoungou 2005: 3-4). In contrast, the country's agroforestry parklands remained a critical source of shea kernels which continued to sustain women's livelihoods by supplying local periodic markets and meeting regional demand for shea butter. Archival records did not enable us to ascertain how out-migration from Upper Volta or corvée may have affected internal labour and market networks.  (Hopkins 1973). Recent research based on estimating the gap between the prices paid to African producers by French trading companies and prices that should have been paid in a counterfactual competitive market showed that African prices were only a small fraction of competitive prices, implying an annual loss of almost 2% of GDP during colonial rule (Tadei 2020). By the outbreak of World War II, their share of total trade increased further, reaching up to 90% for some commodities (Suret-Canale 1971). 25 The colonial trading firm exchange system relied on a regional hierarchical network of local trade posts working with local traders. During this early period, Voltaic wholesalers were not at the head of the regional value chain.

Colonial trading firms
After the economic crises of 1920-21 and 1927-1934, the French colonial administration offered preferential tariffs and incentives to stimulate the production of oilseeds and nuts focusing on groundnuts and copra to reduce France's reliance on foreign, including British, products (Olukoju 2009: 123, Seka 2018). This largely benefited the colonial trading firms and not the farmers. Such protectionism in importing countries and competition from whale oil, Indian Coromandel groundnuts and Sumatran palm produce (which had replaced Nigeria as the world's major exporter of palm oil by 1934, see Martin 2003: 67-68) resulted in negligible interest or investment in shea by the French colonial administration throughout the colonial period. Annual exports of shea nuts and shea butter from AOF were irregular between 1906-1930(Pehaut 1974: 440-447, Figure 3). Haute Volta averaged 670 tons per year prior to the Great Depression, with large annual variations ( Table 2). The large colonial trading firms did, however, start to open new offices in remote markets such as Haute Volta and shea nuts were initially traded by C.F. A.O. in 1939-40 (Coquery-Vidrovitch 1975. The transportation costs from Haute Volta remained high until the Abidjan-Ouagadougou railway opened up in 1955 (Debrie 2010, see also d'Almeida Topor, Chanson-Jabeur andLakroum 1992).

Decline in the shea trade in the late colonial period
Industrialisation was not a priority of GGOAF, particularly in Haute Volta whilst the "pacte colonial" remained in force (Compaore 1984: 11). The first oilseed extraction factory was constructed in 1928 by another colonial trading firm, the Compagnie Française de la Côte d'Ivoire (C.F.C.I.) (Pehaut 1974: 500). It primarily processed groundnuts and exported 1.301 tonnes of oil in 1931. The Director of C.F.C.I. published a review of shea in AOF in which he estimated annual production of shea butter in AOF could reach 15.000 -20.000 tonnes if the "rational culture of shea", originally proposed by Emile Perrot and Jean Vuillet in 1907and 1911respectively, were introduced (Annet 1930: 920). Annet also proposed fixing the purchase price of shea each season through tripartite agreements between local shea producing communities, the colonial administration, and industrial or trading interests (Annet 1930: 920 (CITEC) in Bobo Dioulasso. After 1942, AOF became a vast territory to supply the allies during World War II as the first factories started to transform raw materials into exportable products. In practice, industrialization of the country's two key towns -Ouagadougou and Bobo Dioulasso -only started in earnest after the adoption of the first Investment Code in 1962. The majority of exports from Burkina Faso to the present-day are as unprocessed kernels (AAK is the largest exporter). Limited industrialization took place during the colonial period. Although shea exports did increase in the 1930s, by the time the country gained independence shea production for export had declined again.
As part of preparations to participate in 'the (French) empire and national defense' (Michel 1982), a heavily subsidized 'colonial petroleum' project was developed 27 and was FIGURE 3 Exports of shea nuts and shea butter from Haute Volta, 1901Volta, -1961 Source : Pehaut, 1974 cited in Elias andCarney 2004: 74 26   although the factory did continue to produce biofuels using shea oil for three years between 1950-53 before it was liquidated in 1956 (see also Massa and Madiega 1995: 279-280) ( Table 3). By this stage, metropolitan France had also changed its strategy to diversify its sources of oilseeds. Following an article by Auguste Chevalier (Chevalier 1946), additional research on shea was conducted in the late colonial period (1949)(1950)(1951)(1952)(1953)(1954)(1955)(1956)(1957)(1958) at the Niangoloko research station in the extreme south-west of the country under the auspices of the Institut de Recherches pour les Huiles et O léagineux (IRHO) (see for example, Servant 1956, Ruyssen 1957and Desmarest 1958. This research focused on biological studies and the distribution of shea across the country, propagation techniques and the establishment of shea plantations, and technologies to conserve, dry and extract shea butter. Despite an initiative to develop a new shea research programme (Peperty and Coudeau 1959), IRHO research was suspended before the country gained independence (Picasso 1984). Little additional research on shea was conducted until after the post-2000 boom in the global trade in shea. Picasso noted in 1984 that "in addition to a lack of extension of the first results obtained, the absence of interest and means ensured that little follow-up occurred (Picasso 1984: 43).
In 1956, in an attempt to put an end to marked fluctuations in shea prices in French West Africa due, in part, to the typical three-year cycle in shea production, a first price stabilization fund (Caisse de Stabilisation des Prix) for shea exports was created. Yet, low shea production in the late 1950s counteracted these endeavours and the fund was liquidated in 1959 (Pehaut 1974). Women shea collectors and transformers preferred to use the nuts and processed butter for their own domestic consumption, and/or for sale in local periodic markets and to meet regional (cross-border) demand. A second price stabilization fund Caisse de Stabilisation des Prix des Produits Agricoles) operated between 1968 to 1994 (Government of Burkina Faso/ITC 2015). This is explored in the second paper (Wardell et al. 2021).
After World War II, the colonial state's interests shifted towards groundnuts and cotton as the priority crops to underpin the colony's economic development. Low investment, limited transport infrastructure, changing metropolitan demands for oilseeds, and the ability of local producers to continue to use their own local and regional marketing networks also contributed to the decline of the shea nut export trade in the late colonial period. In essence, the production and supply of shea products remained central to servicing the needs of Burkinabé and West African consumers throughout the colonial period.

OUTCOMES OF COLONIAL EFFORTS TO DEVELOP THE SHEA TRADE
The efforts t o incorporate Haute Volta, a French colonial backwater, into the world economy was drawn out, heterogenous, and messy, and, seen from the perspective of the colonial administration and metropole, ultimately, a failure. The colonial state assumed erroneously that little shea trade existed, and that producers would respond positively to admittedly weak market incentives. We suggest that French colonial policies also failed due to a composite of factors including the limited investment in either the colony or shea as an oilseed crop, the prohibitive costs of road transport and limited transport infrastructure, the adaptation by women shea producers to the extraction of male labour 31 and the trade Sources : Chevalier, 1943: 100;Sissao, 2010: 285;and Pehaut, 1974 1 Supplies of shea nuts were provided to the factory either using forced labour from villages neighbouring the site (until 1946 when force labour was abolished), or by the colonial trading companies which depended on a network of local traders to purchase from local periodic markets up to 150 km from the site. Sissao suggests that SECACO only produced 1.529 tonnes of shea butter from 45.000 tonnes of shea nuts (Sissao, 2010: 283). 28 In metropolitan France, a Comite d'Organisation de carburants et lubrifiants de remplacement was created by decree on 29 September 1942. 29 The isolated site selected was close to the River Mohoun and the then (still) planned extension of the Abidjan-Ouagadougou railway line. The factory only started oil extraction in 1944 but with disappointing results and a low extraction rate (Sissao, 2010). Only a third of the funds released by the Fonds de Solidarite were disbursed in Haute Cote d-Ivoire (Sissao, 2010: 282). 30 This followed additional research conducted by Professor Canac's laboratory at a Scientific Research Centre in Marseille to develop an improved shea butter refining process. AFFECO/904, dossier SECACO. Archives du Centre National d'Outre-Mer, Aix en Provence, France. 31 Haute Volta served as a labour reserve for other colonies in GGAOF. The extraction of male labour left women-headed households with reduced labour for agricultural operations. The collecting, processing and transformation of shea nuts enabled women to spread their labour demand throughout the year to generate both edible oil for their own consumption but as a consistent source of revenues through sales in local periodic markets (Wardell and Fold, 2013).
opportunities created by new international borders, and the 'blindness' of colonial officials to the economic, social and cultural functions of periodic local markets used by women shea traders. Ultimately, the production and supply of shea nuts and shea butter remained central to servicing the needs of Burkinabe and West African consumers throughout the colonial period. Most recent scholarship on the shea trade has focused on relatively recent events and actions, detached from the broader historical context in which it is embedded. This paper emphasizes the emergence of globalized trade which was forged out of distinctive local experiences of production and transformation and local and regional circuits of shea trade controlled by women. By understanding the patterns of pre-colonial and colonial trade in the commodity this helps to explain the gendered nature of the contemporary shea economy in Burkina Faso.

Negligible investment in a French colonial backwater
French West Africa (GGAOF) covered an area nine times the size of France comprising 'coastal' spaces with a population of 5 million which exported 91% of exported products. The other 'continental' spaces including Haute Volta had a total population of 12 million but accounted for less than 10% of agricultural exports (Suret-Canal 1972). The generally low investment in French West Africa, after the 13 April 1900 law removed subsidies to the colonies, was acute in Haute Volta and typified by the belated completion of the railway line linking Abidjan and Ouagadougou in 1955. More than half of the total expenditure on transport infrastructure under the Investment Fund for Economic and Social Development (FIDES) was allocated to Sénégal and Côte d'Ivoire (Debrie 2010: 296). 32 Recent research has highlighted the long-term impacts of early colonial public investments (or lack thereof) in French West Africa (Huillery 2009). Distances to coastal ports and the prices offered for a highvolume low-value commodity such as shea kernels rendered global trade uneconomic. The delays in completing the Abidjan to Ouagadougou railway did not help this situation. Other ambitious, large scale, capital-intensive projects such as the Office du Niger saw Haute Volta primarily as a labour reserve, as was the case with earlier infrastructural and plantation investments in Sénégal and Côte d'Ivoire (Kobi 2019: 47). Colonial efforts to transform indigenous systems of (agricultural) production were, in general, unsuccessful, and probably led to further marginalization of the colony. An estimated 60% of colonies' budgets in French West Africa came from capitation tax. Direct taxes (capitation tax, trading tax and property tax) represented 89% of the total resources of the colonies (Huillery 2009: 181). Local budgets in each colony had to cover all expenses except for military expenses and some of the large-scale public works.
Two initiatives to use shea as a source of gutta-percha and later as a raw material for the production of biodiesel were unsuccessful. The heavily subsidized 'colonial petroleum' project which tried to use shea oil as a feedstock for biofuel production was liquidated in the mid-1950s. A decade of investment in research on shea as a potential oilseed in the late colonial period was abandoned shortly before the country's independence in 1960. Groundnuts and cotton remained the priority export crops in the early and late colonial periods, respectively.

Haute Volta as a labour reserve for other GGAOF colonies
The colonial 'labour reserve' policy in Haute Volta systematically extracted able-bodied men from the territory to provide labour for infrastructural investments in transport infrastructure, cocoa plantations and the timber industry, notably in Côte d'Ivoire and Sénégal. The chiefs ('Chef de Cercle') imposed by the French colonial administration became de facto labour agents for the colonial government and were responsible for encouraging labour migration (Osborn 2003). 33 This left Burkinabe households short of male labour for agricultural production, and heightened the importance to women of value-added processing of shea kernels, and the marketing of kernels and shea butter, all of which they managed themselves. Shea enabled women to spread labour demand for processing and selling shea products throughout the year, and to use the opportunities provided by an elaborate network of periodic markets and trade relays to the south and north of the (new) international boundaries between Burkina Faso and neighbouring countries such as Ghana, Côte d'Ivoire and Benin (Chalfin 2004b, Fold and Reenberg 1999, Wardell and Fold 2013. Hence, the constancy of three-day periodic markets enabled women to sustain their livelihoods and to reproduce social relations devoid of 'boom and bust' cycles, 32 Rene Dumont (1969: 46) noted that "within the framework of FIDES very large sums were granted to French-speaking Africa. In face of the immense needs, however, they seemed quite modest. The aid could in fact have been increased many times without a corresponding tax pressure, had France the courage politically to decolonize more rapidly. Forty-six percent of the FIDES grants, particularly in the first four-year plan, were used to build roads, ports and airports. These were indispensable to open up the countries, but could have been achieved at less cost." In contrast, Paul Nugent (2004) suggests that "the consensus among historians is that FIDES amounted to much more than an ideological fig leaf. It did channel substantial resources into the African colonies -initially (as in the British case) into infrastructural development, but later also into industrial enterprises and agricultural projects." 33 French District Administrators were the real chiefs of the French empire (Delavignette, 1939) and oversaw tax collection, counting people, developing maps, schools, planning and supervising the building of roads, bridges, wells and tracks, arresting criminals and judging them in accordance with the 'native population code' (Code de l'Indigenat). The official tasks of African chiefs were to collect taxes, recruit the workforce for forced labour and recruit military reservists. The amounts to be collected and numbers to recruit were defined by the French administrators (based on Huillery, 2009: 181).
price wars and 'hold-ups' which characterized other globally traded commodities such as palm oil and cocoa in southern Ghana (Milbourn 1970, Lynn 1998. Women were instrumental in sustaining rural livelihoods by using periodic markets on both sides of the border to trade in shea kernels and butter (Fold and Reenberg 1999, Chalfin 2001, Chalfin 2004a, and Wardell and Fold 2013. Hence, colonial labour extraction policies in Haute Volta, mirroring those in the neighbouring Northern Territories of the Gold Coast Colony, may have reinforced the 'gendering' of agroforestry parkland landscapes (Merchant 1990, Norwood 2001, Carney and Elias 2006.

Shea as a peripheral imperial commodity
French tropical agronomy formed part of the "global project for the rational and intensive exploitation of colonial resources" (Sarraut 1923), which specialized in the large-scale production of some priority agricultural products to supply mainland France. Botanical gardens and experimental stations represented the early interface between scientists and colonial administrators. The most important commodities to the French colonial administration, French scientists and French colonial trading firms in French West Africa were groundnuts, palm oil, cocoa, timber and later cotton. Shea was a peripheral 'imperial commodity' throughout the colonial period (Hazareesingh andCurry-Machado 2009, andCurry-Machado 2013) despite the early interest shown in it during the Colonial Exposition in Marseille in 1905 (Henry 1907: 239).
Colonial agricultural development in the GGAOF colonies focused on plant breeding, agronomy, and phytopathology of other prioritized commodities such as groundnuts in Senegal (Bonneuil 1999), cocoa in Côte d'Ivoire (Tourte 2019), and after 1920, cotton in Haute Volta (Tourte 2019c: 431-433, see also . 34 Some efforts were made to rationalize shea production for exports and included proposals to produce latex as a substitute for gutta-percha (Vuillet 1901) and to establish managed forest reserves of shea trees and shea plantations in Katibougou in present-day Mali (Vuillet 1911, Bonneuil and Kleich 2003, and Tourte 2019a. These early initiatives were thwarted, however, as violent resistance to colonial rule culminated in the Volta-Bani War 1915-1916 across Burkina Faso and Mali. 5,000 troops under French authority were deployed to end the anticolonial movement and resulted in the death of an estimated 30,000 Africans, and the destruction of more than a hundred villages (Vennes 2018: 87, Saul andRoyer 2001).
This event triggered the cumulative migration and settlement of an estimated 1.3 million Burkinabe migrants into the neighbouring Protectorate of the Northern Territories of the Gold Coast Colony (Marchal 1999: 222, see also Dresch 1945 andRouch 1956). Migration was to avoid the cumulative burdens of military conscription after 1915 (Echenburg 1975), French capitation tax, forced labour on, for example, collective farms until 1946, 35 and corporal punishment (Asiwaju 1976, Royer 2003. This was compounded by increasingly restrictive land and forest policies that created a 'history of fear' for Burkinabe resource users (Ribot 1995 and1999a).
Research on shea conducted by the Institut de Recherches pour les Huiles et Oléagineux (IRHO) in the late colonial period yielded some promising results before IRHO was suspended in 1958. By this time metropolitan France had changed its strategy to diversify its sources of oilseeds. Little or no investment in additional research on shea was conducted until after the post-2000 boom in the global trade in shea. The volume and value of the shea trade continued to remain insignificant in relation to the dominant French trading firms' interests in groundnuts, timber, and cocoa, and after 1920, the colonial administration's interest in developing cotton production. The volatility of annual shea nut exports varied between 0.5 and 2.9 metric tons during the colonial period. Exports declined in the late 1950s due to over-production of other vegetable oils.

The blindness of colonial officials and scientists
Throughout the colonial era, little effort was made by either the colonial administration or French scientists to understand customary land and resource tenure arrangements, the sociocultural and economic functions of local periodic 3-4 day markets or cross-border trade, and the role of women in either the management of shea parklands or the collection and processing of shea. Colonial officials and scientists alike had little, if any, understanding of either the social functions of periodic markets or the dominant role of women in shea production and marketing processes. In addition, "Colonial officials did not consider the human agency involved in V. paradoxa selection" (Carney and Elias, 2006: 237). The colonial administration and scientists relied on a "world on paper" (Hawkins 2002) determined by maps, laws, decrees and arrêtés, writing a linear concept of history in stark contrast to the knowledge, practices and oral histories of women shea collectors. Local profligate practices including the use of fire as a land management tool were frequently critiqued by colonial officials (see, for example, Laris and Wardell 2006). Women were, furthermore, rarely favoured interlocuters by colonial administrators, scientists or technicians, and the colonial state assumed erroneously that producers would 34 The 20.000 tonnes of cotton set as an annual production objective by Governor Edouard Hesling in 1925 was finally achieved in 1968. This increased to over 400.000 tonnes in the early 2000s (Tourte, 2019c: 433). 35 Collective farms to produce cotton using forced labour were replaced by "model farms" of between 20-60 hectares and by 1930, by "family farms". In 1931, 93 family farms were supervised by the Service de l'Agriculture which, by then, comprised only ten staff. In addition, an experimental cotton farm was established at the Saria Agricultural Station in 1929: by 1934 133 hectares of demonstrations plots of millet, maize, groundnuts and cotton were being used to train young farmers who after two years were expected to establish their own family farms (Tourte, 2019a: 235-237).
respond positively to world market incentives. The French colonial state (and its post-independence successor -see Wardell et al 2021) attempted to regulate indigenous forms of knowledge, gender relations, and social reckoning through legislation. Women shea nut producers and shea nut processing in Haute Volta lived in a "world that didn't count" (McMichael 2004: 7), and remained in the shadow of cocoa growers in Cote d'Ivoire and Ghana until the post-2000 boom in the global shea trade. Women remained, nevertheless, firmly in control of the local and regional shea trade.

The creation of a subversive borderland economy?
We suggest that despite -or perhaps in response to -neglect by the French colonial state, women were instrumental in sustaining rural livelihoods by using local periodic markets to trade in shea kernels and butter on both sides of the new borders separating Haute Volta from its neighbouring territories. These emerged as new international boundaries around the formerly contested Neutral Zone (see Figure 2) were recognized and helped to create 'subversive economies' in the borderlands (Donnan andWilson 1998, see also Wardell et al 2003). The 11 th parallel separating Haute Volta and the Protectorate of the Northern Territories of the Gold Coast Colony has been described as a phantom frontier (Duperray 1999). After the Volta-Bani War, the cumulative exodus of a third of the total estimated Burkinabe population in 1920 (Marchal 1999: 222) and historically embedded patterns of trade and migration in the region (Cordell, Gregory and Piché 1996) helped to create this subversive economy. Ironically, the key factor which influenced the placement of the border separating British and French territories was the protection of metropolitan trading interests. Early attempts by the British to control and regulate the transit trade through the (then) North East Frontier District were abandoned by 1915. 36 In practice, border zones with Ghana, Côte d'Ivoire, Togo and Benin are all distinguished by their ethno-linguistic, cultural, ecological and commercial affinities which has fostered the strengthening of economic alliances and networks outside the formal boundaries of the nation state. State boundaries separating compatible local periodic market systems, differential currency rates and pricing regimes all provided new (and changing) opportunities for women shea nut and shea butter traders to develop and sustain their production and marketing systems to the north and south of the borders. Women also benefited from the immutability of 'marginalized locales' (Chalfin 2001), which remained largely free from both state regulation and the vagaries of global commodity chains.

CONCLUDING REMARKS
West Africa's agroforestry parklands provide a broad range of goods and services at the local, national, and regional levels throughout the region. They constitute part of a complex mosaic of different land uses in landscapes often characterised by many resource users. Their extent and composition have been moulded for centuries by the complex interactions between humans and their biophysical environment. Humans and livestock use land and forest resources, and by doing so they become an integral component of the parkland system, influencing the patterns and processes of change. Humaninduced processes have a strong 'random' element as people tend to adapt continuously to changing conditions and priorities in an unpredictable way (Berry 1993). The shea tree (Vitellaria paradoxa) is still the most commonly occurring species in the parklands of Burkina Faso.
In the early part of the 20th century, the French (and neighbouring British) colonial administrations considered the possibilities of starting large-scale exports of shea kernels to Europe. Early interest focused on the potential use of shea latex as a substitute for gutta-percha. The interest of colonial trading firms led to modest growth in the shea trade in the early colonial period in Haute Volta.
Thereafter, violent resistance to colonial rule during the Volta-Bani War was replaced by persistent migration into the (British) Protectorate of the Northern Territories of the Gold Coast Colony, as a form of protest by local Burkinabé communities to avoid the cumulative burdens of capitation tax, forced labour, military conscription, corporal punishment, and restrictive forest policies. Multiple initiatives to improve the extraction of shea butter (mechanically and chemically), to protect shea parklands or to establish shea plantations, as well as early industrialization efforts, including a French 'colonial petroleum' project, were not successful.
The colonial state assumed erroneously that little shea trade existed, and that producers would respond positively to market incentives. Shea remained a peripheral imperial commodity throughout the colonial era. Women shea collectors and processors played a critical role as agents in this part of the French West African empire as "instigators and promoters" (Curry-Machado 2013: ix) in terms of their capacity to use (or not) the forces of globalization to serve their own interests of local autonomy. French colonial policies also failed due to a composite of factors including the limited investment in either the colony or shea as an oilseed crop, the prohibitive costs of road transport and limited transport infrastructure, the adaptation by women shea producers to the extraction of male labour and the trade opportunities created by new international borders, and the 'blindness' of colonial officials to the economic, social and cultural functions of periodic local markets used by women shea traders.
During the late colonial period, a first attempt was made to put an end to marked fluctuations in shea prices in French West Africa by creating a first price stabilization fund (Caisse de Stabilisation des Prix), but low shea production in the late 1950s counteracted these endeavours and the fund was liquidated in 1959 (Pehaut 1974). Shea collectors and transformers preferred to use the nuts and processed butter for their own domestic food consumption, and/or for sale in local periodic markets including cross-border local markets. Ultimately, the producti on and supply of shea nuts and shea butter remained central to servicing the needs of Burkinabe and West African consumers throughout the colonial period.
Most recent scholarship on the shea trade has focused on relatively recent events and actions, detached from the broader historical context in which it is embedded. Little scholarly work has been published on the history of shea as a colonial commodity. Women were instrumental in sustaining their rural livelihoods by using local periodic markets to trade in shea kernels and butter on both sides of the new borders separating Upper Volta from its neighbouring territories.
The continuities and changes in the governance of shea supply chains in Burkina Faso after independence in 1960 are explored in a second paper 'Shea -the emergence of global production networks, 1960-2021'. By adopting a Global Production Network approach to the evolution of the shea trade in Burkina Faso, the two papers emphasize the emergence of globalized trade which was forged out of distinctive local experiences of production and transformation and local and regional circuits of shea trade controlled by women. Our intention is to demonstrate that one way of dealing with the complexity of globally, regionally and locally traded commodities such as shea nuts and shea butter, is to expand the period under scrutiny to compare the effects of present and previous waves of globalization. The separation of economic processes from social arrangements cannot easily be done, especially not in the longue-durée of socio-economic development.