Non-Timber Forest and Rangeland Products to reduce food insecurity at times of extreme climatic events. A case study in Wedza Communal Area, Zimbabwe

‘Tsubvu’ ( Vitex payos )

Lotte Woittiez

MSc thesis Plant Production Systems Wageningen, March 2010

2

Non-Timber Forest and Rangeland Products to reduce food insecurity at times of extreme climatic events. A case study in Wedza Communal Area, Zimbabwe

Lotte Woittiez

MSc Thesis Plant Production Systems PPS-80433

May 2009 - March 2010

Supervisor: Dr. Mariana Rufino Chairgroup Plant Production Systems Wageningen University Droevendaalsesteeg 1, Wageningen, The Netherlands

Co-supervisor: Dr. Paul Mapfumo Department of Soil Sciences and Agricultural Engineering University of Zimbabwe Mount Pleasant 167, Harare, Zimbabwe

Examiner: Professor Ken Giller Chairgroup Plant Production Systems Wageningen University Droevendaalsesteeg 1, Wageningen, The Netherlands

Dr. Nico de Ridder Chairgroup Plant Production Systems Wageningen University Droevendaalsesteeg 1, Wageningen, The Netherlands

3 4 Table of contents

Preface 7 Summary 9 Chapter 1: Introduction 11 Chapter 2: Methods 13 2.1: Introduction 13 2.2: Definitions 13 2.3: Sampling 13 2.4: Research activities 13 2.5: Data analysis 16 Chapter 3: Study area and sample description 17 3.1: Introduction 17 3.2: Study area description 17 3.3: Sample description 21 Chapter 4: Natural resource availability and access 29 4.1: Introduction 29 4.2: Natural resource availability 29 4.3: Access to natural resources 35 4.4: Discussion 36 Chapter 5: Overview and valuation of NTFRPs 39 5.1: Introduction 39 5.2: Inventory of collected NTFRPs and frequency of use 39 5.3: NTFRP ranking exercise 46 5.4: Discussion 50 Chapter 6: Consumption and collection of key NTFRPs 53 6.1: Introduction 53 6.2: Quantifying NTFRP consumption 53 6.3: Labour for NTFRP collection 56 6.4: Discussion 59 Chapter 7: Diet and the contribution of NTFRPs 61 7.1: Introduction 61 7.2: The contribution of NTFRPs to the diet 61 7.3: Discussion 71 Chapter 8: Livestock management and feed 73 8.1: Introduction 73 8.2: Livestock management in good and bad years 73 8.3: Energy provision for livestock from crop residues and grazing 75 8.4: Discussion 77 Chapter 9: Discussion and conclusions 79 9.1: Discussion 79 9.2: Conclusions 85 References 87 Appendix A: Conversion factors 93 Appendix B: Household questionnaire 97 Appendix C: Access (interview with the headman of Ushe Ward) 107

5 6 Preface

From June to October 2009, I’ve lived the Zimbabwean life: I’ve slept in Zimbabwean houses with Zimbabwean families, I’ve eaten Zimbabwean sadza with muriwo, I’ve walked on the Zimbabwean soil, I’ve sweated under the Zimbabwean sun, I’ve listened to Aleck Macheso (and I’ve danced on it, too) and I’ve met the Zimbabwean people. It was an unforgettable experience, and I want to thank all those kind people that made me feel at home in Zimbabwe. My special thanks goes to Florence and Naboth Mtambanengwe and their entire family, who offered me a home in Harare and a place of belonging. I also want to thank Dr. Paul Mapfumo and Dr. Regis Chikowo from the University of Zimbabwe, who were my supervisors and guides. To Cathrine Mazivanhanga and Tafadzwa: thank you for offering me a home in Ushe and for teaching me how to prepare sadza and muriwo. To Johnson Mupanga: thank you for being my host in Dendenyore and for some nice games of chess. To Magama, Goto, Mr. Chiwaka, MaiMaguenzi and Maimasire: thank you for your invaluable help as translators and guides. To Miss Bake, Shiri, Ernest and my lovely neighbour in Dendenyore whose name I still don’t know: thank you for making me feel at home in rural Zimbabwe. To Hati, Jairos, Grace & Grace, Byron, Christopher, Tony, Tongai and Mr Chitopo: thank you for your help and for the good times that we had. To MaNyashe: thank you for washing the dishes with me. To sensei Joshua: thank you for making my life in Zimbabwe complete by adding a healthy dose of karate. To Manodawafa, Nyamazana, Mandari, Chiwaka, Hakata, Ringoziwa, Gomba, Gware, Mudzungwa, Muza, Kunaka, Tsomondo, Chizavari, Makurumure, Mr. Ushe, Dongo, Phone, Chinhengo, Mawire, Chatukuta, Chakuinga, Madziwa, Tavaziva and Bebulo: thank you for your time, your knowledge, your enthusiasm and your kind hospitality. Without you, this research would have been impossible, and it was an honour to learn from you. To Kurauone and Andy: thank you for being my friend. I would have been lost without you.

Returning to the Netherlands was a shock, and to turn all knowledge and impressions from Zimbabwe into a comprehensible report was not an easy task. Therefore, I want to thank my supervisor, Mariana Rufino, for her guidance and her critical eye. I also want to thank my family and friends for their suggestions and support.

Lotte Woittiez Wageningen, March 10 th 2010.

7 8 Summary

Smallholder farmers in Zimbabwe use the woodlands and rangelands in their community for the collection of products such as firewood, wild fruits and insects, and for the grazing of their livestock. We interviewed 25 farmers, divided over three wealth classes, about the collection and consumption of these so-called NTFRPs (Non-Timber Forest and Rangeland Products) in general and specifically at times of crop failure due to bad weather, usually drought. Farmers were able to name over 130 species of wild plants, insects, mushrooms and animals that were collected from the woodlands and rangelands. The most valued species were Julbernardia globiflora (firewood), Brachystegia spiciformis (firewood), Uapaca kirkiana (fruit), Parinari curatellifolia (fruit), ‘flying termites’ (insect), ‘cape hare’ (animal) and Azanza garckeana (fruit). In good years, households collected on average 4511 kg/year of firewood, 599 kg/year of construction poles, 553 kg/year of leaf litter, 239 kg/year of U. kirkiana , 62 kg/year of P. curatellifolia , 54 kg/year of Strychnos spinosa (fruit) and 36 kg/year of Amanita zambiana (mushroom). In bad years, the consumption of P. curatellifolia increased significantly to 489 kg/year, and the time spent on the collection of both U. kirkiana and P. curatellifolia also significantly increased. For the other products, there were no significant differences between good years and bad years. There was also no significant difference in consumed quantities of any of the products between poor and wealthier farmers, both in good years and in bad years. To look at the NTFRPs in terms of food security, we analysed the contribution of edible NTFRPs to the total energy intake. In good years, all farmers in our sample consumed enough energy to remain above the hunger line, and wealthier farmers consumed more energy than poor farmers. NTFRPs only contributed a minor quantity of the total energy; the major share of energy came from cultivated maize and pulse crops. In bad years, the energy consumption dropped below the hunger line for both the poor and the wealthy farmers in at least some seasons. For wealthier farmers, about 22% of the energy supply came from wild fruits (especially U. kirkiana and P. curatellifolia ) in bad years and for poor farmers, wild fruits supplied up to 42% of the total energy intake. Our data show that wild foods, especially wild fruits, are consumed in large quantities at times of crop failure. Additionally, our analysis suggests that wild fruits contribute greatly to the energy intake of especially poor farmers in bad years. The results of our study support the hypothesis that wild foods can help to reduce food insecurity at times of crop failure due to extreme climatic events.

9 10 Chapter 1: Introduction.

Goal one of the Millennium Development Goals of the United Nations is to eradicate poverty and hunger. The more concrete sub-goal, instead of eradicating hunger, is to halve the ‘proportion of the population below the minimum level of dietary energy consumption’ by 2015 (UN, 2000). Africa, together with Asia and South America, is one of the areas where hunger is still widespread. Most smallholder farmers in sub- Saharan Africa are dependent on crop production and livestock keeping for their food security. In Zimbabwe, smallholder farmers are to a very large extent dependent on rainfall as the sole water source for growing their crops (FAO, 2005). The great majority of the rain in Zimbabwe falls during the rain season, from November to April (Scoones et al. , 1996; Mugabe et al. , 2007). During the rain season, farmers grow maize and other crops such as cowpea, millet, sorghum and sunflower. The whole year round, farmers grow vegetables and sometimes fruits in irrigated gardens, and keep livestock such as cattle, goats and poultry for the provision of animal products and, in case of cattle, for draught power and manure (Ncube et al. , 2009). Livestock graze in the communal rangelands and woodlands and are therefore flexible in their search for different food and water sources (Scoones et al. , 1995).

Most of the smallholder farmers in Zimbabwe live in the so-called Communal Areas, formerly called Reserves or Tribal Trust Lands (O’Flaherty, 1998). In Communal Areas, land tenure takes the form of ‘right of use’, rather than ‘right of property’. Those parts of the land that are not part of a homestead and are not under cultivation belong to all of the community, and can be used for cattle grazing and the harvesting of useful products. Additionally to cropping and livestock keeping, the smallholder farmers in Zimbabwe and other countries of sub-Saharan Africa gather natural products, such as firewood, fruits, insects and medicinal herbs, from the common lands (e.g. Campbell, 1987; Zinyama et al. , 1990; McGregor, 1995; Shackleton and Shackleton, 2002; for a review see Scoones et al. , 1992). These collected products are generally referred to as Non-Timber Forest Products (NTFPs) or as Non-Wood Forest Products (FAO, 2008). A multitude of research projects on the use of non-timber forest products by smallholder farmers in sub-Saharan Africa has been conducted (for an overview see Scoones et al. , 1992; Belcher and Schreckenberg, 2007). Especially the monetary valuation of those products has received a great deal of attention (Peters, 1989; Campbell et al. , 1997; High and Shackleton, 2000; Shackleton et al. , 2002; Dovie, 2003; Dovie et al. , 2007; Kepe, 2007; Belcher and Schreckenberg, 2007). In general, the conducted research projects have led to the conclusion that from a monetary point of view, NTFPs contribute significantly to the livelihoods of smallholder farmers. They can serve as a regular contributor to the overall income or as a safety net to cope with shortages in other livelihood sectors (Zinyama, 1990; Guinand and Lemessa, 2001; Shackleton and Shackleton, 2004; Paumgarten, 2005; Muller and Almedom, 2008).

In the semi-arid climate of Southern Africa, drought is a regularly occurring phenomenon (Le Houerou, 1996; Mazvimavi, 2008). The rainfall patterns are prone to much variation, for example in total amount of rainfall, start of the rains (Philips et al. , 1998) and occurrence of drought spells within the rain season (Usman and Reason, 2004). Rainfall variability can lead to crop failure and food insecurity. Our study analyses the use of NTFPs in relation to climate change or variability, especially drought. The study is part of IDRC project nr 104140 on ‘Exploring measures to

11 enhance the adaptive capacity of local communities to pressures of climate change’, led by the University of Zimbabwe. The first objective of the project is ‘to characterize livelihood profiles of smallholder farming communities in Africa according to their relative capabilities to respond to climate change and variability, paying particular attention to the most vulnerable groups in the context of food security’ (IDRC, 2007). The FAO defines food insecurity as ‘a situation that exists when people lack secure access to sufficient amounts of safe and nutritious food for normal growth and development and an active and healthy life’ (FAO Hunger Portal, 2010). The limit of ‘sufficient’ for Zimbabwe was calculated at a minimum intake of 1800 kcal (7531 kJ) per person per day for the period 2002-2006 (FAO Food Security Indicators Zimbabwe, 2009). According to the definition of the FAO, people who have abundant food in one year but too little in another are still food-insecure. Climate variability, especially drought, often leads to crop failure and thus to food insecurity. The IDRC project objectives target towards increasing the adaptive capacity of smallholder farmers to the possible future decrease of rainfall and increase in frequency of extreme weather situations due to climate change (Hulme et al. , 2001). In our study, we asked the following questions: 1. Can the use of NTFPs be a coping strategy for smallholder farmers at times of extreme weather situations? 2. If yes, then how can NTFPs fulfill this role? 3. How much do NTFPs contribute to household consumption and income? We have the following hypotheses: 1. NTFPs can be a coping strategy for smallholder farmers. 2. Wild foods can prevent or reduce household food insecurity, provided that the resources are accessible for the household. 3. Wild foods contribute relatively more to reducing food insecurity of poor households than of wealthier households because poor households have less assets and are generally less food secure.

In order to test our hypotheses, we quantified the consumption and the collection of NTFPs in good years and in years of extreme weather, using household and key informant interviews, observations and group discussions. Then we analysed the overall household diet and we compared the NTFP contribution to the energy intake in different years and in different wealth groups. In the discussion, we analyse the relevance of our findings and we critically discuss our methods and the possibilities for further research.

12 Chapter 2: Methods.

2.1 Introduction

From July to October 2009, quantitative data on the use of NTFPs was collected during a field study in Wedza district in Zimbabwe. Our main method of data collection was in-depth semi-structured interviewing of a sample of households, along with observations, measurements, group discussions and mapping. The fieldwork was supported by members of the IDRC Climate Change Project team in Zimbabwe, and by agricultural extension workers (AEWs) in the field. The data analysis and report writing were done in the Netherlands, after the data collection was completed.

2.2 Definitions

We use the definition Non-Timber Forest and Rangeland Products (NTFRPs) to describe all products that are collected by the farmers from the common lands, including fallow fields. We include the term ‘rangeland’ because a large share of the products (such as wild vegetables, some mushrooms, termites and animal grazing) comes from fallow fields and areas of veld, rather than from forests. Additionally, from here on, we will use the term ‘woodland’ instead of ‘forest’ whenever the word is not in the phrase ‘Non-Timber Forest and Rangeland Products’, because the term woodland is more correct for the region we describe (Olson, 2001). In order to indicate the differences between weather situations in certain years, we use the terms ‘good year’ and ‘bad year’. The term ‘good year’ describes a year where rainfall is such that an average yield can be achieved. 2008/2009 was a ‘good year’, as well as the years between 2001/2002 and 2006/2007. The term ‘bad year’ describes a year in which the rainfall is such that crops largely fail, usually due to drought. 2007/2008 was a ‘bad year’, as well as 2001/2002 and 1991/1992.

2.3 Sampling

The households in our sample were selected together with the extension workers of each of the respective wards, based on two criteria: location and assets (mainly number of cattle and farm size). Based on asset possession, farmers were divided into three resource groups: RG 1 (wealthy, more than two cattle), RG 2 (medium-wealthy, 1 or 2 cattle or donkeys) and RG 3 (poor, no cattle or donkeys and no additional assets). We selected 12 to 16 households in Ushe (of which in the end 14 were interviewed) and 10 households in Dendenyore (with one more household added on the spot because there was some time left) representing all the different resource groups and divided spatially over the wards.

2.4 Research activities

Household interviews Our main method of data collection was in-depth semi-structured interviewing of a sample of households. All households were visited two to four times, depending on the speed of interviewing and the amount of information required. In each household, the head was asked to cooperate with the interview, and in each case a respondent (either the household head or his wife) offered him/herself voluntarily. During the first interview, a general characterisation of the households was made. Respondents

13 were asked about family size, household assets, farm productivity, input use and yields in the last year (a good year) and in a bad year. Respondents were asked to choose the bad year themselves, based on what they remembered best. Most considered the year 2007-2008 as the year they remembered best for its dramatic weather situation, but some chose the disastrous 1992 season, especially elderly farmers. During the second and third interview, respondents were asked about the collection of NTFRPs from the communal lands. The respondents were requested to list all products that they collected from the common lands and to select the ones that they considered as most important. These ‘most important’ products were discussed further. The respondents were asked to quantify how often a product was consumed or collected, how much was consumed or collected, how many people were involved in the collection, how much labour they spent, how much of the product was kept in store and how much was sold. Additionally, the respondents were asked to explain what the role of a product was, by whom in the family it was used and in what way it was being prepared (if any). In the end of the interviews, a map of the communal areas surrounding the homestead was constructed together with the respondents, and on this map it was indicated from where the different products were collected. Some farmers in Ushe were visited for a third or fourth time because additional information was required. In Dendenyore, because of time limitations and higher efficiency, all farmers were visited only twice.

Key informant interviews In Ushe Ward, the ward headman was interviewed regarding customary laws that govern the use of the communal lands. Other key informants, including extension workers and village heads, were interviewed opportunistically for the purpose of acquiring additional information, for example on access, resource availability and mapping.

Mapping Three different types of maps were constructed: overall ward maps, NTFRP collection maps of individual farmers, and farm maps (only in Dendenyore). The ward maps showed general features (such as roads and rivers) and the natural areas of the ward, divided into veld (grassy areas), bushland (intensively harvested areas with mainly immature trees) and woodland areas (sometimes intensively harvested but with mainly mature trees). The NTFRP collection maps showed the farmer’s homestead and the surrounding natural areas. For each NTFRP that had been discussed, the collection area was indicated on the map. The farm map showed the farmer’s homestead and the fields and garden(s) with the specific sizes, crops and inputs of each.

The overall ward maps were constructed using existing maps, own observations, key informant interviews and group mapping exercises. Existing maps (hand-drawn locally used maps and 1:50000 geographical maps (Surveyor-General Zimbabwe, 1974)) were used to create a framework of ward borders, roads, rivers and mountains. The other methods were used to further fill in this framework. In Ushe Ward, one overall ward map was constructed. For construction of this map, a group mapping exercise was conducted during a meeting of village heads. About 20 of the 29 village heads from Ushe Ward were present and participated in the exercise. The framework-map, copied on an A0-sized sheet of paper, was filled in using stones,

14 maize kernels, sunflower seeds and beans, to indicate the different villages and natural areas (see: Figure 2.1). In the end, one of the participants shaded the indicated areas with coloured pencils. Later, the headman added more details to the map, such as village borders, dip-tanks and churches.

Figure 2.1: Group mapping exercise in Ushe Ward.

In Dendenyore, three different maps were constructed: two in a group-mapping exercise (one by women, one by men) and one by own observations and suggestions of informants. The group mapping exercises were carried out differently than in Ushe. About 15 men and 15 women from a random sample of farmers were given an A0- sized paper with the outline of Dendenyore and the main feature such as roads and rivers. Each group was provided with pencils in three different colours: one for veld, one for bushland and one for woodland. Then, the groups filled in the maps without further guidance. The third map of Dendenyore was constructed using informant interviews and own observations. Informants included extension officers, farmers, village heads and some random by-passers. Own observations were made during the farm visits. For the entire period of interviewing, the map was taken along on every visit and every trip, and thus details were added continuously.

15 2.5 Data analysis

Quantitative data regarding farm production and NTFRP consumption and collection were gathered in several spreadsheets. All statistical analyses were carried out using SPSS 15.0. Differences in collection and consumption patterns between resource groups were analysed using one-way ANOVA with a two-tailed 95% confidence interval of difference. Differences in collection and consumption patterns between good years and bad years were calculated using a paired-samples t-test with a two- tailed 95% confidence interval of difference. Graphs and tables were constructed using Microsoft Office Excel 2003.

16 Chapter 3: Study area and sample description

3.1 Introduction

Zimbabwe is a landlocked country in Southern Africa, with a surface area of 390.580 square kilometers, a length of 852 km WNW-ESE and a width of 710 km NNE-SSW. The study area consisted of two wards located in Wedza district, in the Mashonaland East province in Eastern Zimbabwe (see: Figure 3.1). Wedza district is a former African Reserve, now called ‘Communal Area’. It is inhabited by black smallholder farmers. Land tenure in Zimbabwe is, by and large, divided into two different types: the communal tenure in the former African reserves, and private tenure in the commercial sector (O’Flaherty, 2003). Communal tenure is characterised by a shared property system. The use of land in communal areas is governed by a multitude of rules, that fall under customary as well as under formal law (see paragraph 4.3 and O’Flaherty, 2003). Other than the communal areas, Zimbabwe has resettlement areas that used to belong to white large-scale farmers but that were given back to the Zimbabwean smallholders during the Fast Track Land Reform Programme of the 90’s and 00’s (Moyo, 2005). Here, subsistence farming and small-scale commercial farming systems co-exist. Additionally, some of Zimbabwe’s land is still privately owned and commercially farmed.

3.2 Study area description

Ecology The natural vegetation in Wedza is dry miombo woodland, with Brachystegia boehmii B. spiciformis and Julbernardia globiflora as dominant tree species. Dry miombo woodland stretches over Mozambique, Malawi and Zimbabwe. It is characterised by trees of the typical miombo genera Brachystegia and Julbernardia and by an average rainfall of less than 1000 mm per year (Frost, 1996). The density of the woodland can vary from a closed canopy to a savanna with scattered trees here and there. Fires occur regularly during the dry season (Frost, 1996).

Figure 3.1: Zimbabwe map (a) with the study site (in square), and a satellite image of the region (b) with the location of Ushe Ward and Dendenyore Ward Business Centre, indicated by a pin. Source: Google Maps (a) and Google Earth TM (b) a) b)

17 Site description The field study was carried out in eight villages in Ushe Ward and five villages in Dendenyore Ward. Each ward has an approximate surface of 25 square kilometers which is subdivided into about 30 villages or ‘kraals’. Ushe Ward is located about 20 kilometres South-East of Dendenyore Ward. The Eastern border of Ushe is formed by Ruzawe river, and its Western border is Mhare river, of which a side arm also forms the Northern border. In the South, the border is formed by a range of hills. Throughout Ushe, there run some small streams, several of which are dammed. Most of these streams run almost or completely dry at the end of the dry season. In Dendenyore, water is more plentiful. The Southernmost border is formed by the Save river, which runs all the way to the ocean through Mozambique. In the West run the Jekwa and the Nyamidzi river, the last of which is intensively used for gold panning. The Nyamidzi also forms most of Dendenyore’s Western border. In the East, at least two rivers (Mhare and Nyamemba river) find their source, and each has numerous side-arms. The Eastern border is formed by a range of forested hills. Through all of Dendenyore, there are dams behind which large pools and lakes form, the largest of which is 0.5 km long and almost 0.2 km wide. One of the water sources is utilised in an intensive irrigation scheme, which was constructed by the farmers and improved with help of the European Union. Most of Ushe Ward is relatively flat (apart from the Eastern part) while Dendenyore is mountainous in the South, East and West. Typically, mountains are not cropped (both for reasons of practicality and because of tradition) and therefore most remaining stretches of miombo woodland are found on mountains or in mountainous areas. Dendenyore has relatively large stretches of miombo woodland in the East, the West and all of the Southern part. In Ushe, stretches of undisturbed miombo woodland are found only on the Eastern borders (around the river) and on the scattered mountains. The vegetation type differs between the wards. Notably, the indigenous edible fruit Uapaca kirkiana (wild loquat), which is found mainly in mountainous areas, is scarce in Ushe. In Dendenyore it is very abundant; there are several stretches of woodland consisting almost solely of U. kirkiana in Dendenyore (own observation). The soils in Ushe Ward are fersiallitic sands or sandy loams, with patches of reddish sandy clay loam. The soils in Dendenyore are mostly similar to those in Ushe, but in the North-Eastern part soils are paraferralitic sands and sandy loams (Department of the Surveyor-General, Zimbabwe, 1979).

Climate Even though Ushe and Dendenyore Ward are located less than 25 kilometer apart, they belong to two different agro-ecological zones. Vincent and Thomas (1960) divided Zimbabwe into 5 agro-ecological zones, subdivided into sub-zones. Ushe Ward is located in agro-ecological zone III, which is characterised by a rainfall of 650-800 mm, annually. Dendenyore is located in agro-ecological zone IIb, which is characterised by an annual rainfall of 750-1000 mm. However, over the period 1998- 2007, Ushe received an average annual rainfall of 854±262 mm/year, which is above the long-term average. The difference with Dendenyore (average 873±254 mm/year in 1998-2007) is therefore small. Figure 3.1 shows the monthly rainfall in Ushe (a) and Dendenyore (b) in the last ten years (from satellite data, provided by CIAT). Zimbabwe has suffered several extreme climatic events in the past decades. The 1992 drought is referred to as the worst drought in recent history (Maphosa, 1994; Chenje, 1998). However, the year 2007/2008, which was also a drought year (see: Box 3.1), was extra harsh because of the hyperinflation in Zimbabwe, which caused an overall

18 shortage of food products in the shops. In many places, food was not available for sale at all and farmers had to resort to other sources in order to survive: trading livestock for maize, trading labour for maize, collecting wild foods and getting food from donors and the government. Food aid was donated mainly from October 2008 to March 2009.

Figure 3.1a: Monthly rainfall in Ushe Ward over the last ten years (courtesy of A. Farrow, CIAT). Yearly totals (June - May): ’00/’01 = 1028 mm, ‘01/’02 = 600 mm, ’02/’03 = 669 mm, ’03/’04 = 731 mm, ‘04/’05 = 625 mm, ‘05/’06 = 988 mm, ’06/’07 = 696 mm, ‘07/’08 = 896 mm, ‘08/’09 = 605 mm.

19 Figure 3.1b: Monthly rainfall in Dendenyore Ward over the last ten years (courtesy of A. Farrow, CIAT). Yearly totals (June - May): ’00/’01 = 969 mm, ‘01/’02 = 595 mm, ’02/’03 = 673 mm, ’03/’04 = 754 mm, ‘04/’05 = 768 mm, ‘05/’06 = 1055 mm, ’06/’07 = 771 mm, ‘07/’08 = 929 mm, ‘08/’09 = 643 mm.

20

Box 3.1: Rainfall perception of farmers.

Apart from rainfall quantities from satellite data, we also collected ‘qualitative’ assessments of the climate in good and bad years. For this purpose, farmers were asked to describe the weather situation in the last (‘08/’09) cropping season and in a year that they remembered as being very bad for growing crops. Most farmers described the last cropping season as ‘good’, with ‘fair rainfall distribution’, though some mentioned the occurrence of dry spells or the rain being ‘unreliable’. As year that was very bad for cropping, most farmer chose 2007/2008, because it was a difficult year which was still very fresh in the memory, but some mentioned 2001/2002 and 1991/1992. When asked for a description of the 2007/2008 season, most of the farmers described that in the beginning, there was heavy rain, even with floods, but that in the middle of the season the ‘rain went for good’. The satellite rainfall data (Figure 3.1 a,b) confirm this perception.

3.3 Sample description

Sampling and resource group classification Together with the extension workers, 14 households in Ushe (U) and 11 households in Dendenyore (D) were selected for in-depth interviewing. Table 3.1 shows the assets of each of the households. Based on these assets, nine households were categorised as resource group (RG) 1, eight were categorised as RG 2 and eight were categorised as RG 3. The categories were established according to the classification rules as described in Chapter 2, with two exceptions: household U12 and U14 were placed in RG 2 even though neither owned any cattle or donkeys. Household U12 was categorised as RG 2 because it owned a large area of land (4 hectares of cropping land, 7-8 hectares in total) that was completely fenced. This land resource made the household wealthier than those in RG 3. Household U14 did not ‘own’ any cattle, but the mother-in-law and another family member together owned 5 cattle that were kept in the kraal of household U14. Those animals were available to the household for ploughing and the manure could be used on the fields. Therefore, it was chosen to categorise this household as RG 2. As can be seen in Table 3.1, RG 1 farmers have more cattle than RG 2 and RG 3 farmers. They also have more goats, a bigger farm and a significantly higher maize yield.

Demographic characteristics Table 3.2 shows an overview of the demographic characteristics of the households in our sample. In the RG 1 sample, seven out of nine households were headed by a male. In the RG 3 sample, three out of eight households were headed by a male. With five out of eight households being headed by a male, the RG 2 sample was somewhere in between. Regarding age, the household heads of the RG 1 sample had an average age of 61 and the heads of the RG 2 sample had an average age of 63. However, the heads of the RG 3 sample had an average age of 36, which is significantly lower than the average age of RG 2 and RG 1 heads. As an explanation for this difference, we hypothesise that assets take time to obtain, and that it is therefore likely that more starting, young farmers were classified as RG 3 because they had not yet obtained many assets, whereas older, settled farmers had obtained more assets and therefore were classified as RG 1 or RG 2. Most household heads of the RG 1 group had completed only primary education (seven out of nine). Of the RG 2 group, four out of eight heads had completed only primary education, and two had followed secondary

21 education (O-level). Of the RG 3 group, six out of eight household heads had followed secondary education. Rather than with wealth, this difference likely has to do with age. In 1950, primary education started to be on the rise but there were very few secondary schools for black Africans, and therefore most children finished their education at primary level. From the late 1960s onwards, the number of secondary schools started to increase, with an especially large expansion during the 1980s (WOZA, 2010).

Table 3.1: Assets of sample households. Livestock: cattle = nr of owned cattle and (nr of borrowed cattle), where borrowed cattle is cattle kept in the kraal of the household, but owned by another household or person. Ward Nr RG Farm size Livestock numbers Maize yield Cattle Goats Donkeys Poultry (1/2/3) (ha) (nr (borrowed)) (nr) (nr) (nr) (tonne/yr) Ushe U1 1 2.5 11 5 0 0 2.5 Ushe U2 1 3.0 9 3 0 11 3.5 Ushe U4 1 2.8 5 4 0 6 1.5 Ushe U7 1 5.0 6 1 0 0 1.2 Ushe U8 1 3.2 8 11 0 0 2.3 Dend D1 1 3.4 3 1 0 24 1.8 Dend D7 1 2.3 8 3 0 9 1.5 Dend D8 1 3.0 3 3 0 6 0.8 Dend D10 1 3.8 3 2 0 16 1.0 Average RG 1 3.2 6.2 3.7 0.0 8.0 1.8 Range RG 1 2.5-5.0 3-11 1-11 0 0-24 0.8-3.5 Ushe U3 2 2.0 2 (2) 2 0 14 0.8 Ushe U9 2 2.6 2 0 0 5 1.5 Ushe U10 2 1.2 0 0 2 4 1.2 Ushe U12 2 4.0 0 3 0 9 0.8 Ushe U13 2 1.0 2 2 0 0 0.08 Ushe U14 2 1.6 0 (5) 0 0 12 0.3 Dend D5 2 3.2 2 1 0 16 0.7 Dend D9 2 2.0 0 2 1 10 0.6 Average RG 2 2.2 1.0 (0.9) 1.3 0.4 8.8 0.7 Range RG 2 1.2-4.0 0-2 (0-5) 0-3 0-2 0-16 0.08-1.5 Ushe U5 3 0.6 0 2 0 2 0.6 Ushe U6 3 1.2 0 0 0 4 0.4 Ushe U11 3 2.0 0 0 0 20 0.7 Dend D2 3 2.0 0 1 0 8 0.7 Dend D3 3 2.0 0 0 0 29 0.9 Dend D4 3 1.5 0 2 0 0 1.0 Dend D6 3 2.0 0 0 0 18 1.0 Dend D11 3 0.8 0 0 0 0 0.3 Average RG 3 1.5 0 0.6 0.0 10.1 0.7 Range RG 3 0.6-2.0 0 0-2 0 0-29 0.3-1.0

22 Table 3.2: Demography of sample households. Wards: Ushe= Ushe Ward, Dend = Dendenyore Ward. RG = resource group. Education: P = primary level, J = junior level, O = ordinary level. HH size: adult = age > 18, child = age 3-17, infant = age 0-2. Ward Nr RG Household head HH size Sex Age Educ Adult Child Infant Total (1/2/3) (M/F) (yrs) (*) (nr) (nr) (nr) (nr) Ushe U1 1 M 77 P 3 2 0 5 Ushe U2 1 M 48 P 5 2 1 8 Ushe U4 1 M 67 P 3 3 0 6 Ushe U7 1 M 68 P 2 1 0 3 Ushe U8 1 M 71 P 2 1 0 3 Dend D1 1 F 70 P 2 4 0 6 Dend D7 1 M 46 O 4 2 0 6 Dend D8 1 M 42 P 3 3 0 6 Dend D10 1 F 56 J 2 0 0 2 Average RG 1 7 M, 2 F 61 7 P, 1 J, 1 O 2.9 2.0 0.1 5.0 Range RG 1 42-77 2-5 0-4 0-1 2-8 Ushe U3 2 F 67 2 5 0 7 Ushe U9 2 M 77 P 2 4 0 6 Ushe U10 2 M 58 2 2 0 4 Ushe U12 2 M 55 O 2 0 0 2 Ushe U13 2 F 80 P 2 1 0 3 Ushe U14 2 F 45 P 3 4 0 7 Dend D5 2 M 79 P 5 0 1 6 Dend D9 2 M 40 O 2 2 1 5 Average RG 2 5 M, 3 F 63 4 P, 0 J, 2 O 2.5 2.3 0.3 5.0 Range RG 2 40-80 2-5 0-5 0-1 2-7 Ushe U5 3 F 35 O 1 3 0 4 Ushe U6 3 F 40 O 1 4 0 5 Ushe U11 3 F 41 O 1 3 0 4 Dend D2 3 F 39 O 1 4 0 5 Dend D3 3 M 32 O 2 3 0 5 Dend D4 3 F 45 J 3 3 0 6 Dend D6 3 M 32 P 2 1 0 3 Dend D11 3 M 23 O 2 0 1 3 Average RG 3 3 M, 5 F 36 1 P, 1 J, 6 O 1.6 2.6 0.1 4.4 Range RG 3 23-45 1-3 0-4 0-1 3-6

The average number of adults (older than 17) in RG 1 households was 2.9, with a range of two to five. The average number of adults in RG 3 households was 1.6, with a range of one to three. RG 2 was in between, with an average of 2.5 and a range of 2- 5 adults per household. The number of children (age three to 17) ranged from zero to four or five for each of the resource groups. The number of infants (younger than three) ranged from zero to one for each of the resource groups. Especially in RG 3, there were few households with grown-up children, because the parents were generally still young. Five out of eight households in RG 3 were headed by a single woman. In three out of five cases, the husband had passed away. In one case, the husband worked in town most of the time, and in another, it was unclear whether the husband worked in town to support his family or simply had left.

23 Table 3.3: Income per household per RG. Normal text = information as stated by farmer; text in italic = own observation or stated by AEW; text (between brackets) = unreliable data; * = households with additional income; -- = no data. Income from farming Farm RG Additional sources of income (US $) Good year Good year Bad year

U1 1 0 (NONE) (NONE)

U2* 1 1198 Money from relatives Donor U4* 1 -- Money from relatives Money from relatives U7* 1 -- Money from relatives Donor Donor U8* 1 293 Money from relatives Money from relatives Donor D1* 1 161 Money from relatives Money from relatives D7 1 891 NONE Government 49 D8* 1 Vending Donor + 2 bags of fertiliser D10* 1 20 Money from relatives Money from relatives

U3 2 -- NONE

U9* 2 120 Money from relatives

U10 2 255 NONE Barter trade

U12* 2 -- Arts & crafts Arts & crafts Donor U13* 2 -- Money from relatives Money from relatives U14 2 -- Donor

D5* 2 15 Money from relatives Donor 70 D9* 2 Gold panning Gold panning + 1 bag of fertiliser U5* 3 -- Vending Vending

U6* 3 186 Temporary labour Search for wild food Donor U11 3 -- NONE Barter trade Donor D2* 3 396 Vending Barter trade D3* 3 85 Labourer Gold panning Donor D4* 3 38 Temporary labour Temporary labour D6* 3 205 NONE Labour for food

D11* 3 0 Temporary labour --

24 Income Out of the 25 households in our sample, at least 19 had additional sources of income, apart from the sales of farm produce. The income of each household in our sample is shown in Table 3.3. Of the RG 1 households, at least six out of nine received food and/or money from working relatives. None of the RG 3 households received money from working relatives, but in at least five out of the eight RG 3 households a family member performed temporary labour for cash or food. In bad years, households from all resource groups indicated to receive food aid from a donor, but there were also farmers from RG 1 and RG 2 that indicated they did not receive any food aid because they were ‘too rich’. When looking at Table 3.3, food aid seems to have been given at random; however, the agricultural extension workers (AEWs) told me that there was some kind of selection procedure. The criteria are not clear. Gold panning was a source of income for two farmers in Dendenyore. In Ushe there were no rivers and mountains that were rich in gold. Figure 3.1 shows three gold- panning children.

Figure 3.1: Gold-panning children in Dendenyore.

Household location For each homestead, we did GPS reading to determine the exact location (see: Table 3.4). Figure 3.2 and 3.3 show the location of the different households in Ushe and Dendenyore Ward, respectively. In Ushe, household U1 to U6 were located in the centre of the ward near the tarred road, where the remaining stretches of woodland were few and they were small. Household U7 and U8, as well as household U11 and U12, were located in forested areas between mountains, and these households had relatively undisturbed woodlands nearby. Household U9, U10, U13 and U14 were

25 located in the North of the ward, an area of intermediate density that offered some bushland and woodland nearby. In Dendenyore, household D1 to D3 were located in the most deforested area, close to the secondary school and the business centre along the main dust road. Household D4 to D7 were located close to the range of forested hills on the Eastern border of Dendenyore Ward. Household D8 and D9 were located in an area that was intensively cropped but surrounded by mountains on the West, South and East which provided large stretches of woodland and bushland. Household D10 and D11 were in the middle of the mountains and they were literally surrounded by undisturbed woodlands.

Table 3.4: GPS readings of each household Ward Household GPS reading Ushe U1 18 49 42.9847524 S, 31 49 31.2164104 E Ushe U2 18 49 9.1705448 S, 31 49 23.7307659 E Ushe U3 18 47 44.9781015 S, 31 49 18.1678976 E Ushe U4 18 47 45.3387091 S, 31 49 28.9248851 E Ushe U5 18 49 23.7955469 S, 31 50 8.1430075 E Ushe U6 18 49 18.3503135 S, 31 49 4.1260344 E Ushe U7 18 49 26.9668634 S, 31 52 50.0204285 E Ushe U8 18 49 46.2504896 S, 31 53 10.493912 E Ushe U9 18 46 45.3394163 S, 31 49 58.0986167 E Ushe U10 18 46 18.0621744 S, 31 50 0.7112227 E Ushe U11 18 50 42.6106342 S, 31 51 15.1485753 E Ushe U12 18°50'23.74"S, 31°51'1.92"E Ushe U13 18 46 20.4585144 S, 31 52 0.5021893 E Ushe U14 18 46 15.1965138 S, 31 52 1.8351478 E Dendenyore D1 18 40 47.4157622 S, 31 40 50.8759188 E Dendenyore D2 18 40 49.5222273 S, 31 41 22.2273598 E Dendenyore D3 No data Dendenyore D4 18 43 16.1332082 S, 31 40 49.6210789 E Dendenyore D5 18 44 3.4679505 S, 31 41 8.6811561 E Dendenyore D6 18 42 44.0643818 S, 31 42 9.5842344 E Dendenyore D7 18 42 52.2227072 S, 31 42 3.9231475 E Dendenyore D8 No data Dendenyore D9 18 42 31.4589045 S, 31 39 48.4167581 E Dendenyore D10 18 45 45.9636379 S, 31 40 49.4472764 E Dendenyore D11 18 46 5.5325334 S, 31 40 42.8748295 E

26 Figure 3.2: Satellite image of Ushe Ward with the location of the different households (pins with U-numbers). Red dots = ward border, white/grey = fields, dark green = woodland, red/brown = granite hills/mountains, dotted (as under Madzimbabwe secondary school) = bushland, even green/brown in thin stretches = veld around streams. Source: Google Earth™.

27 Figure 3.3: Satellite image of Dendenyore Ward with the location of the different households (pins with U-numbers). Red dots = ward border, white/grey = fields, dark green = woodland, dotted (as above D2) = bushland, even green/brown in thin stretches = veld around streams. Source: Google Earth™.

28 Chapter 4: Natural resource availability and access

4.1 Introduction

There is a number of studies regarding the use of NTFRPs in Zimbabwe (Chavunduka, 1976; Campbell, 1987; Gomez, 1988; Zinyama, 1990; Grundy et al. , 1993; McGregor, 1995; Campbell et al. , 1997; Kinsey et al. , 1998; Goebel et al. , 2000; Grundy et al. , 2000; Ngwerume and Mvere, 2000; Tyynelä and Niskanen, 2000; Mithofer and Waibel, 2003). However, the subject is by no means exhausted, especially because results from one area cannot simply be extrapolated to all of Zimbabwe. The use of NTFRPs is very dependent on the specific circumstances (such as the resource base) in a certain area. Therefore, we constructed a range of maps to capture the natural resource availability in each of our research areas, as well as the use intensity of and the access to the different natural areas within the community.

4.2 Natural resource availability

Methods Together with the farmers, the community resources were mapped in two different ways, namely by individual mapping and by group mapping. The group mapping procedure has been described in Chapter 2. The individual mapping was carried out at the final stage of each farmers interview. The individual NTFRP collection maps were combined with the large ward maps to construct ‘collection maps’, on which the preferred collection sites were indicated.

Results Figure 4.1 shows an example of an NTFRP collection map of a farmer in Ushe Ward. This map shows the position of the homestead (H), the roads, the river and some areas of woodland and fields. The different products (underlined, with Shona names) have been allocated to the different areas. The number between brackets shows the time that it takes to walk towards the collection place of the product (one-way), according to the respondents. Some products are numbered, in order of preferred collection site. The wild fruit ‘hacha’ ( P. curatellifolia ), for example, is preferably collected close to the homestead, but if there are no fruits available then the farmer walks to tree number two. The map shows a total of 17 NTFRPs that were collected by the respondent. There was an area of woodland available on a nearby mountain, from which the respondent harvested the majority of products: firewood, poles, mushrooms, leaf litter, some herbs and some wild insects. From some areas of woodland further away (one to two hours walking) the respondent collected fruits and mushrooms if those were not available in the woodland nearby, as well as one herb that could not be found elsewhere. Close to home, in the fields, the respondent collected termitaria (soil from termite mounds), termites and thatch grass, and fields a little further away provided grass brooms and P. curatellifolia .

Figure 4.2 (Ushe), and 4.3, 4.4 and 4.5 (Dendenyore) show the results of the group mapping and ward mapping exercises. On the Ushe map, the borders of each village have been indicated, as well as all special buildings such as dip tanks, churches and shops. When comparing the map to a satellite image (see: Figure 3.2) there are a lot of similarities, such as the stretches of woodland along the river in the East, the

29 woodlands on and around the mountains and the bushland/grazing area in the middle, next to Madzimbabwe Secondary School. There are also discrepancies. In the top right of the map, everything is shaded as light-green bushland area, but in fact there are large areas of fields over there, too. Below the orange dot in the top right area (indicated as St Stephen’s Primary School on the satellite image) there should be a mountain which was not on the map. In the middle-bottom of the map, there are no fields indicated, but they should be there. In general, the dark green areas (which are sometimes hard to see on the hand-drawn map) should in many cases have been light- green, because they were more bushlands than woodlands. In bushlands, the canopy were extremely open, there was high bush encroachment and many of the trees were immature because they had been cut down recently.

Figure 4.3, 4.4 and 4.5 from Dendenyore were each constructed in a different way. Figure 4.4 was the result of the men’s group mapping exercise. When comparing this map to the satellite image (Figure 3.3) there are similarities as well as differences. The men’s group has drawn mainly large natural areas of one type or the other, but in fact the satellite image shows that the landscape is much more fragmented. Only along the Eastern border and on the mountain range in the West, large stretches of woodland are found. In the middle, all natural areas are fragmented by fields. Figure 4.5 was the result of the women’s group mapping exercise. For some reason, the women chose to work with squares, instead of with shapes that were representative for the actual areas. In general there are few resemblances between the satellite image and the women’s map, but the women did place the names of all the villages of Dendenyore on the map, at their supposed location. Figure 4.3 was constructed bit by bit, by own observations and the inputs of a range of informants, including the respondents from the sample households. The map shows mainly the areas of collection that were important to the respondents. Large areas that none of the respondents used are missing, such as the woodland on the mountain range in the South-West.

The black dots on Figure 4.2 and 4.3 show the position of the different households in our sample, and the white dots show the collection sites of each NTFRP that was discussed. In Ushe Ward, most farmers appear to stay relatively close to their homestead when collecting NTFRPs; the white dots are clearly concentrated around the black dots. The majority of the NTFRPs in Ushe was collected from woodlands and bushlands, but considerable numbers were also collected from the fields. For some NTFRPs, such as certain insects, certain mushrooms, fruits and thatch grass, fields were the most abundant collection areas. However, other NTFRPs such as firewood, leaf litter and most mushrooms could be found only or mainly in woodlands. For two farmers, firewood was collected from field areas because no bushland and woodland areas were available nearby, and several farmers had to walk far because the nearby areas of woodland or bushland were already stripped of all the dead wood. This was especially the case in the Western part of Ushe. At least three farmers complained that they had to walk very far for collecting mushrooms, because there were no forested mountains nearby. Collecting U. kirkiana was a very time- consuming activity for most of the Ushe farmers, because they had to walk about 20 kilometers to the West, all the way to the Eastern border of Dendenyore, to find U. kirkiana in abundance.

30 Figure 4.1: Farmers’ NTFRP collection map.

Figure 4.2: Ushe Ward map combined with individual NTFRP collection maps. The black dots represent the households and the white dots represent the different NTFRPs, at the location where they are collected. Colour key: red = tarred road, blue = stream/river, yellow = dust road, light green shading = bushland, dark green shading = woodland.

31 Figure 4.3: Dendenyore Ward map (from own observations and informant contributions) combined with individual NTFRP collection maps. The black dots represent the households and the white dots represent the different NTFRPs, at the location where they are collected. Colour key: red = tarred road, blue = stream/river, brown = dust road, light green shading = bushland, dark green shading = woodland, orange shading = sourveld, yellow shading = sweetveld, purple shading = fields, grey shading = gardens.

32 Figure 4.4: Dendenyore Ward map, constructed by the men’s group during the group mapping exercise. Colour key: red = tarred road, blue = stream/river, brown = dust road, light green shading = woodland, pink shading = bushland, brown shading = veld.

33 Figure 4.5: Dendenyore Ward map, constructed by the women’s group during the group mapping exercise. Colour key: red = tarred road, blue = stream/river, brown = dust road, dark green squares = woodland, orange squares = bushland, grey squares = veld.

34

In Dendenyore Ward, none of the farmers complained about having to collect firewood from the fields; in fact, for each farmer there were woodlands nearby that could be used for firewood harvesting. For the wild fruits, especially U. kirkiana and P. curatellifolia , farmers usually walked to the woodlands in the Eastern half of the ward, or on the Eastern border. For cattle grazing, the areas of veld and wetland along the many streams were intensively utilised. In general, farmers in Dendenyore collected only specific products such as grasshoppers, termites and certain wild vegetables from the fields. For all other products, specific areas such as woodlands or velds were available.

4.3 Access to natural resources

Methods The presence of natural resources does not necessarily mean that these resources are available to all community members. In order to complete the picture of natural resource availability, we analysed the access rules in Ushe and Dendenyore Ward. The respondents were first asked to indicate natural areas of limited access within their communities, but no such areas existed according to the respondents. Then, the respondents were asked to indicate on the map to which areas they could or could not go for collecting firewood and wild fruits, hunting wild animals or grazing their cattle. The answers are shown in Table 4.1. In order to triangulate the answers of the respondents, we held an interview with the headman of Ushe Ward regarding access and usage rules for the communal lands. The results of this interview are summarised below.

Results: access to natural resources according to the respondents Table 4.1 shows the rules for the collection of four different NTFRP types, according to the respondents. The answers that respondents gave ranged from ‘no collection at all is allowed’ to ‘you can go anywhere you want to collect this’, and anything in between. Some respondents gave an answer that fit none of the categories, and these answers were classified as ‘other’. Regarding the grazing of cattle, most respondents (eight out of 14) thought that they could let their cattle graze only within their ward, and two respondents even stated that cattle could graze only within the village borders. Regarding the collection of firewood, half of the respondents (eight out of 16) said that they had to stay within village borders, but two others said that they could go anywhere, even outside the ward, to collect firewood. One respondent said that firewood could be collected only in the area that was appointed by the headman. For the collection of wild fruits, the respondents were rather unanimous: 13 out of 15 said that wild fruits could be collected anywhere, regardless of the ward borders. For wild animals, there were only few respondents that answered the question, but of those respondents half thought that the animals could be hunted anywhere, one said that animals could only be hunted within village borders and one said that hunting was allowed anywhere, but setting traps was only allowed within village borders.

35 Table 4.1: Collection and usage rights as indicated by farmers. ‘Total’ (last column) shows the total nr of respondents that answered the question regarding access. In and around Type Own village village Own ward Anywhere None at all Other Total Grazing 2 3 3 5 0 1 14 Firewood 8 2 3 2 0 1 16 Fruits 0 0 2 13 0 0 15 Animals 1 0 1 3 1 0 6

Results: access natural resources according to the headman of Ushe Ward Mr. Ushe, the headman of Ushe Ward, had been in office for two years at the time of the interview. According to Mr. Ushe, the custom was that indigenous foods, especially plants, were available to anyone who needed them. During the 2007/2008 drought, the collection of Parinari curatellifolia and Uapaca kirkiana fruits in other wards was negotiated with the villagers there. In some cases, when people had trees in their yard, these trees could be only picked after asking permission, but according to the customary law this permission could not be denied if the fruit tree was indigenous because indigenous fruit trees cannot be 'owned' by someone. It was not allowed, according to customary law, to sell any indigenous fruits. Dried firewood, either wood that was already on the ground or dead branches that were still on the tree, could be collected freely by kraal members within the kraal boundaries. To collect firewood in another kraal, one had to ask permission from the kraalhead. This was common practice, as some kraals had much more woodland than others. For cutting trees, permission was required from the kraalhead. Trees on the mountains could never be cut, for reasons of erosion, security (e.g. as a hiding place in events of war) and tradition. The mountains were for example used in the rituals of spirit mediums and the trees on the mountains were the domain of the ancestor spirits and had to be preserved. According to customary law, all animals could be hunted freely. Hunting in other wards was not allowed. According to national law, no hunting at all was allowed. In the past, cattle grazing was restricted to grazing areas. Nowadays, cattle could graze in most of the ward, ignoring kraal boundaries. The kraalhead could appoint some areas that were prohibited for cattle grazing, and fields under cropping were always prohibited. It was not allowed to graze cattle in other wards, but in very special cases, permission could be asked from the ward headman. Kraalheads controlled the obedience to customary law within their kraal. Kraal boundaries coincided with natural boundaries such as streams, low-lying wetland regions and mountains. The headman held court every Friday, and if cases of illegal use of natural resources came up the headman could demand a fine in the form of cash, a chicken or a goat. For the entire interview, see: Appendix C.

4.4 Discussion

The maps resulting from the group mapping exercises give some very accurate and some less accurate representations of the actual vegetation. Both on the Ushe and the Dendenyore maps, details were missing, such as areas of cultivation within stretches of woodland or bushland. The Ushe map was constructed by village heads, and they were generally very knowledgeable about their own village. However, the group was very large and therefore it was not always possible to capture all the knowledge from each of the participants. The group mapping in Dendenyore was done by a mixture of

36 ‘normal’ farmers and village heads for the men’s group, and only ‘normal’ farmers for the women’s group. The men’s group was more capable of producing an accurate map with the different natural areas. The fact that the women were able to allocate all the villages to their (more or less) correct position in the ward shows that they were able to read the map, but to draw the natural areas was apparently a very difficult assignment. In general, the sheer size of the area to map (the surface of Dendenyore is > 25 square kilometers) made it difficult to capture detail. A village-by-village group mapping exercise would likely result in a more accurate, more detailed map. Both in Ushe and in Dendenyore Ward, woodland and bushland areas were utilised for the collection of NTFRPs. In Ushe, there were some areas where NTFRP availability was limited because the woodlands or bushlands were far away or intensively harvested. Whenever possible, the respondents collected the NTFRPs in an area close to the homestead. Sometimes respondents had to collect NTFRPs from fields or degraded bushlands in which the availability was very low. In Dendenyore, natural areas were more abundant and the respondents selected the most suitable natural areas for the collection of specific NTFRPs. Both for Ushe and for Dendenyore, the respondents mentioned that the availability of most of the NTFRPs was declining, because woodlands and trees disappeared. The interview with the headman from Ushe Ward yielded a number of clear customary laws. However, these laws were apparently not so clear for the respondents. In some cases, the differences between the theoretical rules as described by the headman, and the actual rules as applied by the farmers, may simply have been a matter of application. For example, the headman stated that the collection of wild fruits from other wards in 2007/2008 was negotiated, but for the farmers this just meant that they could go to another ward to collect fruits and that people from other wards could come over and do the same. The collection of firewood in another village may not be allowed without permission, but for those farmers that had been given permission in the past it was a normal thing to do. Still, the difference between ‘theory’ and ‘practice’ does not explain everything. For example, also in 2009, when there was no food shortage, farmers from Ushe went to Dendenyore to collect U. kirkiana , which is illegal according to the headman. To find out why the rules were not clear to the farmers would be a study in itself. Nevertheless, it is something to keep in mind when considering the sustainability of the use of the communal lands. Without a clear set of rules, it is unlikely that NTFRPs are, or will be, sustainably harvested, which means that the resource availability is likely to decline.

37 38 Chapter 5: Overview and valuation of NTFRPs

5.1 Introduction

In order to fully capture the use of NTFRPs in our specific study area, we started with making an inventory of the total range of NTFRPs that were collected by the farmers in our sample. To compose this list, at the end of the first or at the beginning of the second interview the respondents were asked to name all the species or products that they collected within the following categories: fruits, vegetables, roots/tubers, herbs, mushrooms, animals, insects, feed and non-food. The categories had been established on beforehand, based on literature (Gomez, 1988; McGregor, 1995). Additionally, we asked the respondents to make a selection of NTFRPs that they valued most, and to estimate the relative importance of the different NTFRPs at times of good yields as well as at times of crop failure.

5.2 Inventory of collected NTFRPs and frequency of use

Methods We composed a list of all the species and products that the respondents could name. Per product p we scored for each household h within a resource group r whether it did (k = 1) or did not ( k = 0) collect the product. Per product, we then summed the k- scores per resource group to calculate the collection score N (see: equation 5.1). The results are shown in Table 5.1. The collection score (A) for all households, in the last column of Table 5.1, was calculated by summing the results of all the resource groups (see: equation 5.2). The overall collection score per product category c is shown at the bottom of each sub-table per resource group r (T, see: equation 5.3) and for the entire sample ( TA , see: equation 5.4).

Eq. 5.1: Collection score of a product per resource group.

n r = r N p ∑k ,hp , h=1 where N is the collection score of product p per resource group r, and k is the collection coefficient of product p for household h ( h = 1, 2, ..., n) in resource group r. If product p is collected by household h in resource group r, then k = 1, otherwise k = 0.

Eq. 5.2 : Collection score of a product for all households.

n = r Ap ∑ N p , r =1 where A is the overall collection score of product p, summed up over all households.

39 Eq. 5.3 : Collection score of a product category per resource group.

n ,cr = ,cr T ∑ N p , p=1 where T is the collection score of product category c per resource group r. Product categories are fruits, vegetables, herbs, roots/tubers, mushrooms, insects, animals and non-food.

Eq. 5.4 : Collection score of a certain product category for all households.

n TA c = ∑T ,cr , r =1 where TA is the overall collection score of product category c, summed up over all households.

Results The complete list of collected products and species, and the collection scores per product or species, are shown in Table 5.1. All households indicated to extract firewood for cooking and heating. None of the households had an electricity connection, and no other alternative energy sources were available for these purposes, either. Ninety-two percent of the households indicated to collect the mushroom Amanita zambiana which was described as ‘a good relish’ by several respondents. Eighty-eight percent and eighty-four percent of the households indicated to collect the wild fruits Parinari curatellifolia and Uapaca kirkiana , respectively. Grazing of animals was only included in the questionnaire in Dendenyore Ward, and ten out of the eleven interviewed households (91%) indicated to use the common lands for grazing their livestock, either cattle, donkeys, goats or poultry. The eleventh household did not own any livestock at the time of the interview.

A total of twelve species of food products was mentioned by more than half of the farmers (13 or more households). This total included four species of fruit ( U. kirkiana, P. curatellifolia, Strychnos spinosa and Azanza garckeana ), three species of mushroom ( A. zambiana , ‘tsvukesvuke’ and Boletus edulis ), three species of insect (winged termites, soldier termites and Christmas beetle), one species of vegetable (Corchorus oditorius ) and one species of herb ( Lippia javanica ). Six types of non- food products (firewood, termitaria, leaf litter, poles, thatch and grazing) were collected by more than half of the households, but this number should probably be higher; grass brooms, for example, could be found in every household that was interviewed, but apparently farmers easily forgot to mention them. The preferred firewood species were Julbernardia globiflora and Brachystegia spiciformis .

40 Table 5.1a: Wild fruits, species names and collection scores. FRUITS Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Parinari curatellifolia Mobola plum Hacha 8 7 7 22 Uapaca kirkiana Wild loquat Mazhanje 7 7 7 21 Strychnos spinosa Bitter monkey-orange Matamba 5 6 6 17 Azanza garckeana Snot apple Matohwe 6 3 5 14 Vitex payos Chocolate berry Tsubvu 5 5 2 12 Dovyalis caffra Kol apple Nhunguru 7 3 2 12 Ficus sycamorus Fig Mawonde 2 4 4 10 Vanqueriopsis lanciflora Crooked false medlar Matufu 4 3 2 9 Syzygium guineense Hute 2 3 3 8 Lannea edulis Tsambatsi 2 3 3 8 Ximenia caffra Sour plum Tsvanzva 3 3 2 8 Garcinia huilensis Matunduru 2 2 2 6 Carissa edulis Carissa Dzambiringwa 1 1 3 5 Anona senegalensis Custard apple Maroro 2 2 0 4 Tsokotsiana 1 1 2 4 Strychnos innocua Monkey orange Makwakwa 0 1 1 2 Carissa bispinosa Carissa Munzambara 1 0 1 2 Sclerocarya caffra Marula Marula 1 1 0 2 Tsvirinzvi 0 1 0 1 Diospyros mespiliformis Ebony Shuma 1 0 0 1 Bridelia cathartica Mupambare 1 0 0 1 Carica papaya Pawpaw Mapopo 0 1 0 1 Dovyalis caffra Kei apple Tsvoritoto 0 0 1 1 Ficus burkei Wild fig Tsamvi 0 0 1 1 Psidium guajava Guava Guava 0 0 1 1 Adansonia digitata Baobab Mahuyu 0 1 0 1 Opuntia vulgaris Prickly pear Zvinanazi 0 1 0 1 Masadzambodza 0 1 0 1 Piliostigma thonningii African biscuit Musekesa 0 0 1 1 Ximenia caffra Sour plum Nhengeni 0 1 0 1 Mimusops zeyheri Red milkwood Chechete 0 1 0 1 Total fruits 61 62 56 179

41 Table 5.1b: Wild vegetables, species names and collection scores. VEGETABLES Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Corchorus oditorius Bush okra Derere 6 3 5 14 Gynandropsis gynandra African spider herb Nyeve 2 2 3 7 Amaranthus spp. Poor man's spinach Mhowa 2 2 3 7 Senecio erubescens Chirewerewe 4 2 1 7 Sesamum angustifolium Sesame Samuwende 2 2 1 5 Cleome monophylla Spindle pod Mujakari 1 2 1 4 Heteropogan contortus Spear grass Mhuvuyu/mutsine 1 1 2 4 Mhonja 2 0 0 2 Solanum nigrum Nightshade Musungusungu 1 1 0 2 Monenza 0 1 0 1 Fototo 0 1 0 1 Mundya 0 1 0 1 Chipesu 0 1 0 1 Total vegetables 21 19 16 56

Table 5.1c: Wild herbs and medicines, species names and collection scores. HERBS Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Lippia javanica Lippia/menthe Zumbani 4 6 5 15 Temnocalyx obovatus Makoni tea bush Makoni tea 2 0 3 5 Aloe spp. Aloe Gawagawa 3 0 2 5 Dicoma anomala Chifumuro 1 1 3 5 Erythrina abyssinica Lucky-bean tree Mutiti 2 0 1 3 Elephantorrhiza elephantina Muzezepasi 2 0 1 3 Moringa 1 0 1 2 Eucalyptus spp. Gum tree 1 0 1 2 Terminalia sericea Mususu 1 0 1 2 Sarcostemma viminale Milk rope Nyokadombo 2 0 0 2 Muwengahonye 0 1 1 2 Manyama 0 0 1 1 Securidaca longepedunculata Violet tree Mufufu 0 0 1 1 Christmas tree Christmas tree 0 0 1 1 Guava Guava coffee 0 0 1 1 Ficus sycamorus Fig Muwonde leaves 0 0 1 1 Cyperus angolensis White-flowered sedge Chityorabadza 0 0 1 1 Combretum apiculatum Mugodo 1 0 0 1 Gardenrule 0 1 0 1 Solanum incanum Bitter apple Nhundurwa 1 0 0 1 Munzvanzva 1 0 0 1 Mumhungu 1 0 0 1 Mutarara 1 0 0 1 Muroro 1 0 0 1 Ndolani 0 1 0 1 Lantana camara Lantana camara 0 1 0 1 Musahute 0 1 0 1 Total herbs 25 12 25 62

42 Table 5.1d: Wild roots/tubers, species names and collection scores. ROOTS/TUBERS Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Coleus esculentus Vlei tuber Tsenza 3 1 2 6 Eriosema pauciflorum Tsombori 1 1 3 5 Babyana hypogaea Hwenya 1 1 0 2 Commiphora marlothii Paperbark Munyera 0 1 0 1 Muchanya 1 0 0 1 Total tubers 6 4 5 15

Table 5.1e: Wild mushrooms, species names and collection scores. MUSHROOMS Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Amanita zambiana Zambian slender caesar Nhedzi 9 6 8 23 Cantharellus Tsvuketsvuke 9 4 6 19 Boletus edulis Tindindi 6 2 6 14 Cantharellus Chihombiro 4 3 4 11 Cantharellus densifolius Nzeve(ambuya) 6 3 3 12 Termitomycete Huvhe 5 1 2 8 Uzutwe 1 2 3 6 Ndebvudzasekuru 0 2 1 3 Bunaretsoko 1 2 0 3 Chinyokashesheshe 1 1 0 2 Dindijava 0 2 0 2 "Mushrooms" Owa 0 2 0 2 Chiyambwe 0 0 1 1 Tsihhuri 0 1 0 1 Chiropachembwa 0 0 1 1 Total mushrooms 42 31 35 108

Table 5.1f: Insects, species names and collection scores. INSECTS Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Macrotermes spp. Flying termites Ishwa 8 5 5 18 Macrotermes spp. Soldier termites Majuru 7 3 4 14 Eulepida masnona Christmas beetle Mandere 5 4 4 13 Orthoptera spp Grasshoppers/locusts Whiza/mashu 4 2 5 11 Coimbrasia belina Caterpillars Madora 2 3 5 10 Carebara vidua Flying ants Tsambarafuta 3 2 3 8 Bracytypus membranaceus Makurwe 3 0 3 6 Cirina forda Harati 2 0 1 3 Sternocera funebris Zvigakata 1 0 1 2 Total insects 35 19 31 85

43 Table 5.1g: Wild animals, species names and collection scores. ANIMALS Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Sylvicapra grimmia Common duiker Membwe 4 4 4 12 Lepus capensis Cape hare Tsuro 5 2 4 11 Mice Mbeva 4 1 3 8 Birds Shiri 2 1 4 7 Procavia capensis Rock rabbit Mbira 2 3 1 6 Potomachoerus larvatus Wild pig Nguruve 2 2 2 6 Fish Hove 1 0 2 3 Paracynictis selousi Selous mongoose Jerenyenje 0 2 1 3 Hystrix africaeausralis Porcupine Nungu 0 1 1 2 Numida meleagris Wild guineafowl Hanga 0 1 1 2 Paraxerus cecapi Tree squirrel Tsindi 0 0 1 1 Fowl Orwe 0 0 1 1 Oreotragus oreotragus Klipspringer Ngururu 0 1 0 1 Nhimba 0 1 0 1 Chiwuta 0 1 0 1 Aepyceros melampus Impala Impala 0 0 1 1 Raphicerus campestris Steenbok Mhene 0 1 0 1 Total animals 20 21 26 67

Table 5.1h: Livestock feed, species names and collection scores. LIVESTOCK FEED Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 Piliostigma thonningii African biscuit Musekesa 4 4 0 8 Mpangara 3 1 1 5 Tsokotsiana 0 0 1 1 Muhunga 1 0 0 1 Pfubvudza 1 0 0 1 Star grass 1 0 0 1 Green glass 1 0 0 1 Total feed 11 5 2 18

44 Table 5.1i: Non-food products, species names and collection scores. NON-FOOD Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 FIREWOOD Firewood Huni 9 8 8 25 Julbernardia globiflora Munhondo 5 5 4 14 Brachystegia spiciformis Msasa 6 4 4 14 Brachystegia glaucescens Mountain acacia Muwunze 3 4 2 9 Brachystegia boehmii Mupfuti 2 6 1 9 Combretum apiculatum Mugodo 2 1 3 Pericopsis angolensis Muwanga 1 1 2 Piliostigma thonningii Musekesa/mutukutu 1 1 2 Mudjoke 1 1 2 Dovyalis caffra Kol apple Munhunguru 1 1 Mudzunzowa 1 1 Azanza garckeana Mutohwe 1 1 Eucalyptus spp Gum tree Eucalyptus 1 1 Mushawa 1 1 Dichrostachys cinerea Mupangara 1 1 Mubuku 1 1 Tsokotsiana 1 1 TERMITARIA Termitaria Churu 7 5 8 20 LEAF LITTER Leaf litter Mutsakwani 8 5 7 20 Julbernardia globiflora Munhondo 1 1 2 Brachystegia boehmii Mupfuti 1 1 2 Ficus burkei/ingens/natalensis Wild fig Mutsamvi 1 1 2 Brachystegia glaucescens Mountain acacia Muwunze 2 2 Sectia brachypetala Mutondochuru 1 1 Ficus sycamorus Fig Muwonde 1 1 Piliostigma thonningii Musekesa 1 1 Ziziphus mucronata Muchecheni 1 1 Brachystegia spiciformis Msasa 1 1 Tsokotsiana 1 1 Mukonachando 1 1 POLES Poles Mapango 6 6 6 18 Eucalyptus spp Gum tree Eucalyptus 1 1 1 3 Pericopsis angolensis Muwanga 2 1 3 Terminalia sericea Mususu 1 1 2 Burkea africana Mukarati 2 2 Ormocarpum trichocarpum Mpotanzou 1 1 Mubuku 1 1 Murwiti 1 1 THATCH Thatch grass Huswa 5 2 7 14 Hyparrhenia filipendula Madangaruswa 1 1 Nutu 1 1

45

NON-FOOD (continued) Latin name English name Shona name Nr of times mentioned per RG RG 1 RG 2 RG 3 ALL n=9 n=8 n=8 n=25 GRAZING Grazing 4 2 4 10* ROPE Rope from bark Makavi 1 0 3 4 Brachystegia spiciformis Msasa 1 1 BROOMS Brooms 1 2 1 4 Myrothamnus flabellifolius Mufandichimuka 1 1 Schotia brachypetala Mwawashuni 1 1 STONES Stones for building 0 0 1 1 SODIC SOIL Sodic soil 1 0 0 1 BRICKS Bricks 0 1 0 1 Total non-food 42 31 45 118 * Grazing was scored only in Dendenyore Ward.

Honey and birds eggs, as well as wooden tools and utensils, were also not mentioned by any farmers, though previous research projects have found that these are important NTFRPs (Campbell et al. , 1997; Shackleton et al. , 2002). The above tables must therefore be seen as an indication of the most important NTFRPs found in Wedza, rather than as an exhaustive list.

5.3 NTFRP ranking exercise

Methods After listing the NTFRPs that were collected by the household, respondents were asked to choose five NTFRPs that they normally valued most, and to put those five products in order of importance. Then, they were asked to do the same, but now for a bad year. The ranks were then converted to values as follows: rank 1 = value 5; rank 2 = 4; rank 3 = 3; rank 4 = 2; rank 5 = 1. Some farmers picked 5 products without hierarchy; these were awarded a value of 3, each. If farmers had ranked a product group (e.g. ‘mushrooms’) instead of a single species, the highest ranking product of that group was given that value.

The values ( v) were used to calculate the relative importance of each product (see: Table 5.2). The total sum of values ( V) per product p per resource group r was divided over the total sum of values that was awarded by farmers in each ward ( Vw ) to get a weighted value ( WV , see: equation 5.5 – 5.7). This weighted value was corrected because the non-food products had only been included in the ranking exercise in Dendenyore, and therefore their relative value was calculated based on the sum of values for Dendenyore only ( Vw D), which gave an unbalance in the calculation. By multiplying with a specific correction factor ( C), the unbalance was solved (see: equation 5.8 and 5.9). In the end the numbers were multiplied by 100 to get readable results (see: Table 5.2). In Box 5.1, the calculations are illustrated with two examples.

46 Eq. 5.5: Value of a product per resource group in a certain year.

n r = r V , yp ∑v ,, yhp , h=1 where V is the value of product p for resource group r in year y, and v is the value of product p for household h ( h = 1, 2, ..., n) in year y. Year y is either good or bad.

Eq. 5.6: Total of all values awarded in a resource group within a ward.

m n ,rw = ,rw Vw y ∑∑v ,, yhp , h=1p = 1 where Vw is the sum of the values of all products p (p = 1, 2, ..., n) for all households h (h = 1, 2, ..., n) in resource group r in ward w in year y. The wards are Ushe ( U) and Dendenyore ( D).

Eq. 5.7: Weighted value of a food or a non-food product per resource group.

 V r = ⇒ r = , yp  pt F WV , yp n  ,rw  ∑Vw y  w=1 ,  V r  = ⇒ r = , yp pt NF WV , yp ,rD  Vw y where pt is the product type, either food ( F) or non-food ( NF ), and WV is the weighted value of product p for resource group r in year y. Vw D is the sum of the values of all products p for all households h in resource group r in Dendenyore Ward.

Eq. 5.8: The resource-group specific correction factor.

n r = r Cy ∑WV , yp , p=1

where C is the correction factor for resource group r in year y.

Eq. 5.9 : The relative value of a product per resource group.

WV r r = , yp RV , yp r *100 , Cy

where RV is the relative value of product p for resource group r in year y.

47 Box 5.1: Example calculations of equation 5.5 to 5.9. Example 1 (food product) Example 2 (non-food product) p = P. curatellifolia (pa) p = firewood (fi) r = resource group 1 (1) r = resource group 3 (3) y = good year (g) y = good year (g)

1 3 Eq. 5.5 V ,gpa = 3 V ,gfi = 1+1+5+1+5 = 13 U 1, U 3, Eq. 5.6 Vw g = 69 Vw g = 30 D 1, D 3, Vw g = 60 Vw g = 75 3 13 Eq. 5.7 WV 1 = = 0.023 WV 3 = = 0.17 ,gpa 69 + 60 ,gfi 75 1 3 Eq. 5.8 Cg = 1.16 Cg = 1.10 023.0 17.0 Eq. 5.9 RV 1 = *100 = 2.0 RV 3 = *100 = 15.7 ,gpa 1.16 ,gfi 1.10

Table 5.2: Total relative value per product per resource group in good and bad years. GOOD YEAR BAD YEAR Product name Relative value per product per RG Relative value per product per RG RG 1 RG 2 RG 3 Total RG 1 RG 2 RG 3 Total FRUITS Parinari curatellifolia 2 3 0 2 27 31 22 26 Uapaca kirkiana 10 7 14 10 8 20 10 12 Strychnos spinosa 2 2 3 2 2 5 2 3 Azanza garckeana 5 5 3 4 0 0 0 0 Vitex payos 1 4 0 2 5 4 0 3 Dovyalis caffra 3 0 0 1 0 0 0 0 Ficus sycamorus 0 0 4 1 3 0 0 1 Vanqueriopsis lanciflora 4 0 0 2 3 0 0 1 Syzygium guineense 3 0 0 1 0 4 0 1 Carissa edulis 0 0 0 0 0 0 3 1 Strychnos innocua 0 0 0 0 0 0 3 1 Carissa bispinosa 1 0 0 1 0 0 0 0 Sclerocarya caffra 0 0 0 0 0 2 0 1 Psidium guajava 0 0 0 0 0 0 1 0 Fruits TOTAL 30 22 23 26 47 65 42 50 VEGETABLES Corchorus oditorius 0 0 4 1 4 0 1 2 Gynandropsis gynandra 0 2 0 1 2 0 0 1 Amaranthus spp 0 0 3 1 0 0 0 0 Senecio erubescens 3 0 0 1 2 0 0 1 ‘Mhuvuyu’ 0 0 4 1 0 0 0 0 ‘Vegetables’ 0 0 0 0 0 0 4 1 Vegetables TOTAL 3 2 12 6 8 0 5 5

48 Table 5.2, continued GOOD YEAR BAD YEAR Product name Relative value per product per RG Relative value per product per RG RG 1 RG 2 RG 3 Total RG 1 RG 2 RG 3 Total HERBS & MEDICINES Aloe spp 0 0 3 1 1 0 0 0 Dicoma anomala 0 2 7 3 0 0 2 1 Erythrina abyssinica 1 0 0 0 0 0 0 0 Ficus sycamorus leaves 0 0 0 0 0 0 1 0 ‘Medicine’ 0 3 0 1 0 4 0 1 Herbs TOTAL 1 6 9 5 1 4 3 2 ROOTS/TUBERS Coleus esculentus 0 0 0 0 0 0 7 2 Eriosema pauciflorum 0 0 0 0 3 0 0 1 Roots/tubers TOTAL 0 0 0 0 3 0 7 4 MUSHROOMS Amanita zambiana 2 4 4 3 4 4 0 3 ‘Tsvuketsvuke’ 0 0 0 0 2 0 0 1 Boletus edulis 0 0 3 1 0 0 0 0 ‘Chihombiro’ 0 2 0 1 0 0 0 0 ‘Huvhe’ 3 0 0 1 0 0 0 0 ‘Uzutwe’ 0 0 3 1 0 0 3 1 ‘Bunaretsoko’ 0 0 0 0 1 0 0 0 ‘Mushrooms’ 3 4 0 2 0 6 0 2 Mushrooms TOTAL 8 10 9 9 6 10 3 6 INSECTS Flying termites 13 1 0 6 2 0 2 2 Soldier termites 3 0 0 1 2 4 0 2 Eulepida masnona 0 0 0 0 1 0 3 1 Grasshoppers 3 0 3 2 4 2 3 3 Coimbrasia belina 0 0 0 0 0 0 3 1 Insects TOTAL 19 1 3 9 8 6 11 9 ANIMALS Common duiker 0 3 3 2 0 0 4 1 Cape hare 7 4 0 4 3 0 4 3 ‘Mice’ 1 0 0 0 1 0 0 0 ‘Birds’ 3 3 3 3 0 0 1 0 Wild pig 0 0 0 0 2 0 0 1 ‘Fish’ 3 0 3 2 0 0 0 0 Animals TOTAL 14 11 9 12 6 0 9 6 NON-FOOD Firewood 14 23 16 17 13 16 0 9 Leaf litter 0 0 0 0 0 0 11 4 Poles 0 16 7 7 0 0 0 0 Thatch 4 9 4 5 6 0 0 2 Grazing 6 0 0 2 1 0 8 4 Rope 1 0 0 1 0 0 0 0 Stones 0 0 6 3 0 0 0 0 Non-food TOTAL 26 48 33 34 21 16 19 19

49 Results The relative values per NTFRP are shown in Table 5.2. In good years, a total relative value of 34% was given to non-food products, with firewood being the most highly valued product (17%), followed by poles (7%) and thatch grass (5%). Wild fruits followed the non-food items as most highly valued product type, with a total relative value of 26% being awarded to this category. Uapaca kirkiana stood out as the most highly valued food product (10%) followed by flying termites (6%), Cape hare (4%), Azanza garckeana (4%), Dicoma anomala (3%), Amanita zambiana (3%) and ‘birds’ (3%). When we look at the number of times that the different products or species were mentioned (Table 5.1) and at the value that was attributed to them (Table 5.2) there were clear differences. Frequency of collection and attributed value are thus not necessarily related. The best example is cape hare, which was collected/hunted by only 44% of the households but was highly valued by those, ending as third most highly valued food product in our list.

In good years, RG 1 farmers valued wild fruits, insects and wild animals relatively highly, but they did not attach much value to wild herbs and wild roots/tubers. The most-valued products for RG 1 farmers were the wild fruits, which were valued even higher than non-food products (including firewood). The RG 2 farmers valued non- food products especially highly. In good years, they awarded 48% of the total value for NTFRPs to non-food products. They did not attach much value to insects, but wild animals and especially wild fruits were considered relatively important. Unfortunately, the valuation results of RG 2 are based on a very small sample because some farmers did not do the exercise and some did not collect more than two or three products. As a result, the valuations of two farmers skewed the overall RG 2 valuation towards the non-food products. For RG 3 farmers, non-food products were also most important (33% of the total) but wild fruits, wild vegetables and wild herbs were valued highly as well. Farmers from all resource groups attached a value of about 10% of the total to wild mushrooms. Wild roots/tubers were important for none of the resource groups in good years.

In bad years, wild fruits were the most highly valued category of NTFRPs, with an overall relative value of 50%. Valuation of the non-food products dropped to 19% of the total. Parinari curatellifolia became the most highly valued product, with a relative value that went from 2% in good years to 26% in bad years. Firewood was valued at only 9% of the total, and the values of thatch grass and poles also decreased, but grazing became higher valued, probably because there was little other feed available for the livestock. Valuation of roots and tubers became higher, but mushrooms were considered less important because of their limited availability in times of drought. Regarding the animals, several farmers mentioned that they did not hunt in years of drought because of low availability and competition of other households.

5.4 Discussion

The respondents were able to name over 130 NTFRP species that were regularly collected. Especially for wild fruits and wild herbs, farmers were able to mention a large range of species. For most categories, the RG 1 farmers appeared to use the widest range of products. This likely had to do with age. RG 1 households were generally led by an elderly head, whereas especially RG 3 households were often

50 young families or middle-aged single mothers with children (see: Table 3.2). Elderly farmers had more traditional knowledge on the use of wild plant and animals, and they appeared to be better able to identify different species. Younger farmers were less knowledgeable and also seemed to be less attached to the consumption of traditional gathered products. Not all of the products that were collected by the respondents were considered as ‘very important’. Over half of all the species mentioned was considered as ‘one of the five most important NTFRPs’ by none of the farmers. On the other hand, some products, such as cape hare, were collected by less than half of the farmers but were nevertheless valued highly. It must be noted that the sensitivity of the valuations was high. A single farmer with a great enthusiasm for a specific species could highly affect the overall outcome of the valuation exercise. There were some remarkable results. For example, insect and animal products were valued much higher by RG 1 than by RG 3 farmers. However, in general RG 3 farmers had much less animal proteins in their diet since they owned less livestock. Therefore, one would expect them to value animal products that can be collected from the common lands relatively high. But this was not the case. Also, RG 1 farmers valued wild fruits higher than RG 3 farmers did. Again, this is remarkable because RG 1 farmers grew more exotic fruits than RG 3 farmers, so it would make sense if RG 3 farmers relied more on wild fruits and valued these products higher. Instead, RG 3 farmers appeared to find wild vegetables, herbs and non-food products relatively important. As an explanation, one could offer that animal products and fruits are luxury products, whereas vegetables (eaten with maize porridge) and non-food products are more basic. From this perspective, it makes sense that the wealthier farmers attached more value to the fruits and animal products whereas poor farmers found basic products more important.

The increased valuation of wild foods (especially wild fruits) in bad years supports the notion that in such years, wild foods play a role in preventing or relieving food insecurity and hunger. However, the ranking does not support the notion that resource poor farmers are more reliant on wild foods than wealthier farmers. Especially wild fruits, which serve as snacks in times of good yield but as meals in times of shortage, were valued more highly by wealthy than by poor farmers.

51 52 Chapter 6: Consumption and collection of key NTFRPs

6.1 Introduction

In the first part of the farmer’s interviews, we made an elaborate farm characterisation and also we composed a list of NTFRPs that the farmers collected from the common lands. Furthermore, we attached a relative value to these NTFRPs by asking the respondents to select the products they found most important, and to give them a rank. Thus, we answered the question which NTFRPs were collected in our study area, and how important these NTFRPs were in good years and in bad years. However, we did not yet get any quantitative information on the NTFRP collection and use. The second part of the interviews was focused on collecting this quantitative information.

6.2 Quantifying NTFRP consumption

Methods After the ranking exercise, a subset of NTFRPs was selected for further analysis, based on the respondent’s ranking and the researchers’ consideration. For each selected product, the respondents were asked how often they consumed the product and how much they consumed per consumption event, both in good and bad years. The respondents were also asked about the number of collection events, the quantities collected and the destination of the collected product: either consumption, storage, or sales. Finally, the respondents were asked to further specify the use of the products, to indicate substitutes and to assess whether the availability of the product had changed over time. To assess the reliability of the respondents’ estimates regarding the quantities of use, they were asked to estimate consumed and collected quantities, separately (note: for some of the non-food products, only collection was quantified because ‘consumption’ was not relevant). In some cases, there was a factor 2 up to a factor 10 difference between the respondents’ estimate of quantities consumed and quantities collected. Therefore, we analysed the reliability of our data by testing for a significant difference between the ‘consumed’ and ‘collected’ quantities per product. The results are shown below. The respondents generally answered the quantitative questions in local units, such as ‘one bucket’, ‘two cups’ or ‘a handful’. Therefore, the answers needed to be translated to standard values (grams or kilograms) before any further calculations could be done, which required a number of conversion coefficients and assumptions (see: Appendix A). Out of the total of 55 products that were discussed in detail with one or more respondents, we selected the 15 NTFRPs that were overall valued highest (see: Table 5.2). We limited ourselves to these fifteen because for the next products in rank, too little information was available to draw any statistically sound conclusions. However, there were no products of the categories ‘vegetables’ and ‘roots/tubers’ among the 15 highest ranking products. Therefore, we decided to add Corchorus oditorius and Coleus esculentus to the selection, which were the highest ranking products in the ‘vegetables’ and the ‘roots/tubers’ category, respectively. Thus, we came to a selection of 17 key NTFRPs on which we did further analysis. To calculate the average used quantities (see Table 6.1, column ‘average’ (in kg fresh weight per year)), the sum of used quantities was averaged over the number of households (n) that provided data for the product. For our calculations, we used the ‘consumed’ quantities unless only ‘collected’ quantities were available or unless ‘collected’

53 quantities were lower than consumed quantities. Our reason (based on our observations) was that households generally collected their own NTFRPs and thus, that they could not consume what they did not collect. This way, we also prevented overestimation of the importance of the NTFRPs.

Results Table 6.1a shows the collected quantities of the key NTFRPs, collected in good years. Firewood is on top of the list, with collected quantities of 4511 kg/household/year, on average. Because firewood is used every day and is available all year round, in contrast to for example the fruits and mushrooms, which are usually strictly seasonal, it is according to expectations that firewood is collected in the largest quantities. On the bottom of the list is Dicoma anomala , a herb known also as ‘chifumuro’ or ‘maagbitterwortel’, with an average consumed quantity of 200 g/hh/yr. In good years, Uapaca kirkiana was the most-consumed food product, with an average consumed quantity of 239 kg/hh/yr. To put this in perspective: the staple, maize, was consumed at quantities of approximately 1000 kg/hh/year. The U. kirkiana fruit was highly appreciated by the respondents because of its sweet taste. Farmers from Ushe walked long distances (up to 20 km there and 20 km back) to collect it, whereas in Dendenyore the mountains were literally covered with U. kirkiana trees. The second most consumed food was the fruit P. curatellifolia , followed closely by the fruit S. spinosa . Fruits are easy to collect and when a mature tree is near, they come in large quantities. This may explain why the three most collected food products were all fruits. The fourth most collected food, A. zambiana , was a mushroom. Mushrooms are available only when there is rain, but at these periods they come in large quantities so farmers tended to collect them with baskets full.

In bad years (Table 6.1b) the consumption of P. curatellifolia significantly increased from 62 kg/hh/year to 489 kg/hh/year, on average over all resource groups. Consumption of U. kirkiana and S. spinosa also greatly increased, but these results were not statistically significant. In fact, the consumption of all edible NTFRPs, apart from A. zambiana , flying termites and birds, increased in bad years. A. zambiana and flying termite consumption necessarily went down, because both products are only available after rain. Birds were hardly collected at any time, but according to the respondents the availability was greatly reduced in years of drought. The total consumption of wild foods in years of drought easily surpassed the consumption of cultivated foods. Average maize yields dropped to less than 500 kg/year, and of the other crops only the vegetables (irrigated by hand) yielded reasonably well (see: Table 7.2). Thus, P. curatellifolia and U. kirkiana alone provided as much harvest, in terms of amounts, as the field crops did. These data show that wild foods, especially wild fruits, greatly contribute to the diet in bad years.

There were no significant differences in consumed quantities between the different resource groups for any of the products, either in good or in bad years. In fact we sometimes found large differences indeed, but non of these was significant and additionally, there was no clear trend. In good years, for example, RG 1 farmers collected most P. curatellifolia , A. zambiana , Corchorus oditorius and cape hare, RG 2 farmers collected most poles, thatch grass, Azanza garckeana and grasshoppers and RG 3 farmers collected most firewood, leaf litter, U. kirkiana and Coleus esculentus . In bad years, especially the RG 3 farmers appeared to collect larger quantities of wild

54 fruits and RG 1 farmers appeared to collect relatively less of all. But again, these results were not significant.

Table 6.1: Average consumption (kgFW/household/year) of key NTFRPs in a) good years and b) bad years. Stars (*) indicate significant differences between good and bad years per product. Differences between good and bad years per resource group were tested using a paired-sample t- test. Differences between resource groups were tested using one-way ANOVA. a) GOOD YEAR Product RG 1 RG 2 RG 3 TOTAL Average Std n Average Std n Average Std n Average Std n (kg/hh/yr) (kg/hh/yr) (kg/hh/yr) (kg/hh/yr) Firewood 4736 2657 9 2938 790 8 5829 3372 8 4511 2697 25 Poles for construction 388 18 2 788 796 5 425 214 3 599 576 10 Leaf litter 294 84 4 328 298 3 839 675 6 553 531 13 Uapaca kirkiana 146 130 7 258 367 5 332 395 6 239 301 18 Parinari curatellifolia 149 387 8 16 31 6 2.8 7.4 7 62* 240 21 Strychnos spinosa 39 1 53 22 3 73 1 54 20 5 Amanita zambiana 65 47 4 19 13 4 20 17 3 36 36 11 Thatch grass 33 7.7 3 60 36 2 10 14 2 34 26 7 Corchorus oditorius 39 55 2 0 2.3 1 27 44 3 Cape hare 35 28 2 0 12 0.0 2 23 21 4 Flying termites 14 15 6 52 73 2 6.3 1 21 34 9 Vitex payos 4.0 3.0 2 23 33 3 0 15 25 5 Azanza garckeana 0.06 1 16 22 2 0.5 1 8.3 16 4 Coleus esculentus 0 0 6.1 2.5 2 6.1 2.5 2 Grasshoppers 2.7 2.3 2 3.6 1 3.0 2.6 3 3.0 2.0 6 Birds 0 2.3 1 2.3 1 2.3 0.0 2 Dicoma anomala 0 0.002 1 0 0 2 0.0 0.0 3

b) BAD YEAR Product RG 1 RG 2 RG 3 TOTAL Average Std n Average Std n Average Std n Average Std n (kg/hh/yr) (kg/hh/yr) (kg/hh/yr) (kg/hh/yr) Firewood 4736 2657 9 2938 790 8 5829 3372 8 4511 2697 25 Poles for construction 388 18 2 788 796 5 425 214 3 599 576 10 Leaf litter 294 84 4 610 614 4 716 675 5 553 531 13 Uapaca kirkiana 377 281 7 455 352 5 1008 1200 6 609 752 18 Parinari curatellifolia 434 449 8 560 428 7 481 444 7 489* 423 22 Strychnos spinosa 243 1 97 53 3 592 322 2 286 285 6 Amanita zambiana 24 31 4 18 15 4 42 67 3 27 37 11 Thatch grass 33 7.7 3 60 36 2 10 14 2 34 26 7 Corchorus oditorius 78 110 2 0 0 78 110 2 Cape hare 31 33 2 0 62 14 2 46 27 4 Flying termites 2.2 3.1 6 62 84 2 2.7 1 15 40 9 Vitex payos 1.0 0.2 2 64 74 3 0 39 63 5 Azanza garckeana 2.7 1 19 18 2 3.0 1 11 14 4 Coleus esculentus 0 0 239 91 2 239 91 2 Grasshoppers 0.5 0.8 2 25 1 0.2 0.3 2 5.4 11 5 Birds 0 2.3 1 0.3 1 1.3 1.4 2 Dicoma anomala 0 1 0.3 0.4 2 0.2 0.3 3

55 6.3 Labour for NTFRP collection

Introduction For smallholder farmers in Southern Africa, labour is generally a constraining factor in crop production (Alwang and Siegel, 1999; Zingore et al. , 2009). Therefore, the amount of labour that is allocated to the production or collection of a product tells something about the importance of this product for the household. Mithöfer and Waibel (2003) studied the investment of labour by farmers in Murehwa district, Zimbabwe. They found that per household, about 600 hours/year were invested in agriculture, another 600 hrs/year in horticulture and about 1700 hrs/yr were invested in livestock keeping.

Methods To estimate the labour investment in the collection of NTFRPs, respondents were asked to estimate how much time they spent on collecting the most relevant NTFRPs, and to indicate which household members were responsible for the collection. These data were used to calculate the total labour allocation per year per product per household, in good and in bad years. Averages were calculated over those households that contributed data (see: Table 6.2). The significance of the difference between labour allocation in good and bad years was calculated using a two-tailed paired sample t-test.

Results The products in Table 6.2 are ordered as in Table 6.1, according to the total collected amounts in good years. The biggest outlier were the birds, which required large time investments for small collected quantities. In fact, for all the insect and animal products, labour requirements appeared to be relatively high, either because the animals need to be hunted or because the collection was carried out by children and involved a lot of playing around. Second-most labour was invested in the collection of firewood, which was also collected in the largest quantities. On average, farmers spent 477 hours per year on the collecting firewood. Additionally, U. kirkiana , S. spinosa and A. zambiana took a large share of time, with a labour investment of between 124 and 84 hours per year for each.

In bad years, farmers invested most of their time in the collection of firewood, P. curatellifolia , U. kirkiana and Corchorus oditorius , but also in the hunting of cape hare and birds. Corchorus oditorius and the birds will not be discussed further, because the data are based on a single household and therefore they are not very representative. Whereas P. curatellifolia and U. kirkiana were consumed in high quantities as meals, hare meat was consumed only in low quantities (on average 46 kg/hh/yr), which did not really warrant the large time investment. For U. kirkiana and P. curatellifolia , labour investment in the collection increased significantly in bad years compared to good years when looking at the entire sample. For U. kirkiana , the average labour investment per year went from 124 hrs/yr in good years to 377 hrs/yr in bad years. For P. curatellifolia , labour investment went from 17 to 236 hrs/yr. For none of the other products, there was a significant difference in labour investment between good and bad years.

56 Table 6.2: Labour allocation per key NTFRP in a) good years and b) bad years. In the last column, significant differences in labour allocation between good and bad years are indicated (*= p<0.05; **= p<0.01). Differences between good and bad years were tested using paired-sample analysis. LABOUR ALLOCATION TO NTFRP COLLECTION GOOD YEAR BAD YEAR Significance Product n Total labour (hr/hh/yr) n Total labour (hr/hh/yr) Firewood 25 477 24 535 ns Poles for construction 4 39 4 39 ns Leaf litter 13 52 9 52 ns Uapaca kirkiana 13 124 16 377 * Parinari curatellifolia 6 17 20 236 ** Strychnos spinosa 4 87 4 116 ns Amanita zambiana 7 84 5 51 ns Thatch grass 3 54 3 54 ns Corchorus oditorius 3 9 1 832 - Cape hare 4 109 4 454 ns Flying termites 8 96 6 0 ns Vitex payos 4 39 3 39 ns Azanza garckeana 1 10 - Coleus esculentus 2 12 2 61 ns Grasshoppers 5 205 2 105 ns Birds 2 1016 1 576 - Dicoma anomala 2 1 1 0 -

Labour allocation can be based either on collected quantities, where farmers collect large quantities of a certain product and therefore invest much labour, or on the value of a product for the household. If quantities are the only rationale, then one would expect a clear positive correlation between labour investment and collected quantities. In good years, when all data are included, this correlation was negligible (r = 0.313), but when the birds were excluded from the calculation, there was a significant positive correlation (r = 0.87, p < 0.001) between labour investment and collected quantities. In bad years, when all data were included, there was no correlation between collection time and collected quantities (r = 0.377). When the birds and Corchorus oditorius , which were both outliers with values based on data from only one farmer, were excluded from the sample, there was a small but significant positive correlation (r = 0.65, p < 0.05). The major outliers that remained were cape hare and grasshoppers, which required large time investments to catch small quantities. The reverse goes for Coleus esculentus , a tuber, for which large quantities could be collected in relatively little time.

One may expect that competition and depletion in bad years reduce labour efficiency (collected quantities per hour of invested labour), but this does not appear to be the case. For four of the food products, including P. curatellifolia and U. kirkiana , labour efficiency was indeed reduced in bad years, but for six others, including S. spinosa and A. zambiana , efficiency of collection actually increased. Intensification of collection may be the cause of this; S. spinosa for example was collected per piece in good years, but per bucket in bad years, leading to a reduction in time investment per collected kilogram.

57 Only the labour investment in the collection of P. curatellifolia and U. kirkiana increased significantly in bad years. Table 6.3 shows the labour investment per resource group for these two fruits, in good and bad years. We found no significant differences in labour investment between the resource groups, and only for U. kirkiana did we find a significant difference (p < 0.05) within a resource group, namely RG 1. RG 1 farmers invested significantly more labour in the collection of U. kirkiana in bad years than in good years. Both for U. kirkiana and P. curatellifolia , the labour investment appears to run more or less parallel with the consumed quantities (see: Table 6.1 and 6.3).

Table 6.3: Labour allocation for the collection of Uapaca kirkiana and Parinari curatellifolia per resource group. Significant differences (p < 0.05) are indicated with a star (*). Product Weather RG1 RG2 RG3 Average (hr/yr) n Average (hr/yr) n Average (hr/yr) n Uapaca kirkiana Good 67* 6 208 4 124 5 Bad 230* 7 437 5 501 6 Parinari curatellifolia Good 42 7 4.4 5 0.57 7 Bad 212 7 307 7 188 7

The labour division in our sample over males and females, and over adults and children, is shown in Table 6.4. The male-female division seems rather even, but with only males hunting the animals (cape hare) and only females collecting roots (Corchorus oditorius ) and thatch grass. Both males and females participated in the collection of firewood, fruits, insects and mushrooms. Furthermore, only the collection of birds was left to children alone. For all other products, adults also participated in the collection, though grasshoppers, Vitex payos , Strychnos spinosa and P. curatellifolia were mainly collected by children.

Table 6.4: Gender and age division in the collection of key NTFRPs. AGE AND GENDER-BASED DIVISION OF LABOUR Age division Gender division Fraction of adult Fraction of child Fraction of male Fraction of Product n labour labour labour female labour Firewood 25 0.59 0.41 0.41 0.59 Poles for construction 4 0.36 0.64 0.82 0.18 Leaf litter 13 0.58 0.42 0.50 0.50 Uapaca kirkiana 13 0.20 0.80 0.58 0.42 Parinari curatellifolia 6 0.62 0.38 0.50 0.50 Strychnos spinosa 4 0.11 0.89 0.54 0.46 Amanita zambiana 7 0.48 0.52 0.41 0.59 Thatch grass 3 1.00 0.00 0.00 1.00 Corchorus oditorius 3 1.00 0.00 0.00 1.00 Cape hare 4 0.67 0.33 1.00 0.00 Flying termites 8 0.38 0.62 0.72 0.28 Vitex payos 4 0.25 0.75 0.45 0.55 Grasshoppers 5 0.23 0.77 0.46 0.54

58 6.4 Discussion

Neither in good nor in bad years, there was a significant difference in the consumption of the key NTFRPs between the resource groups. Also, for none of the individual resource groups, consumption level of an NTFRP increased or decreased significantly in years of bad weather. However, when looking at the entire sample, the consumption of P. curatellifolia was significantly higher in bad years than in good years. When looking at the consumed quantities of all wild fruits in bad years, the data suggest that RG 3 farmers consumed more P. curatellifolia and S. spinosa than RG 1 farmers, but these differences were not significant. It is important to note that the consumed amounts per NTFRP cannot be extrapolated to the entire population because the average is calculated based only on those households that provided data for the product, and those were only the households that found the product important. Households that did not at all use a product or that found a product not important were selectively left out. Therefore, the sample was not random and the consumption of some of the key NTFRPs is very likely to be an overestimation. However, the overall consumption of NTFRPs per household was probably an underestimation because only five to ten out of all the consumed NTFRPs per household were discussed and quantified while many more were actually consumed, in amounts that we do not know.

Because the respondents were sometimes inconsistent in their estimates of ‘consumed’ and ‘collected’ quantities of NTFRPs, we analysed if there were structural differences. For firewood, average ‘collected’ quantities were significantly higher than average ‘consumed’ quantities (two-tailed student t-test, p < 0.05), but this was probably caused by the conversion of consumption (estimated in bundles or piles) and collection (estimated in scotch carts or bundles) to kilograms. We assume that the weight of a cartload of firewood was overestimated, and therefore the collected quantities were structurally too high (see: Appendix A). Firewood was the only NTFRP for which there was a significant difference between ‘consumed’ and ‘collected’ quantities. For all other NTFRPs, there were variations but they were not significant.

Labour availability is limited, and therefore the investment of labour can indicate the importance of a product or activity. Our data show that in bad years, average labour investment of all farmers in the collection of P. curatellifolia and U. kirkiana significantly increased. The consumed amounts of P. curatellifolia in bad years were also significantly higher. There were no other significant changes in consumed quantities or labour investment between the years. Our data also show that labour investment by RG 1 farmers in the collection of U. kirkiana but not P. curatellifolia increased significantly, but that the consumption of U. kirkiana did not increase significantly in bad years. There was no significant increase of consumption or labour investment within any other resource group for any other NTFRP, and there were no significant differences at all between resource groups. Thus, our results show that on average, farmers consume significantly more of the wild fruit P. curatellifolia and invest significantly more time in the collection of P. curatellifolia and U. kirkiana in bad years. These conclusions are in agreement with our hypothesis that in bad years, farmers collect more wild foods to relieve food insufficiency. However, the data do not support the hypothesis that poor farmers are more dependent on NTFRPs than rich farmers. Even though the RG 3 farmers appear

59 to consume larger quantities of P. curatellifolia and S. spinosa than the other farmers in bad years, these results are not significant. In fact, our data seem to support the opposite of our hypothesis, because only the wealthy RG 1 farmers invest significantly more times in the collection of Uapaca kirkiana in bad years, which indicates their dependence on wild foods at times of crop failure.

60 Chapter 7: Diet and the contribution of NTFRPs

7.1 Introduction

According to the FAO Food Security Indicators for Zimbabwe (2009), an average Zimbabwean needs to consume at least 1800 kcal (7531 kJ) per day in order not to be ‘hungry’. This comes down to about 229 MJ per month, or 2749 MJ/year. Maize is the main energy supplier, with an average energy content of 15.2 MJ per kilogram dry maize (FAO, 1968). However, our data show that NTFRPs are also regularly consumed in large quantities. In this chapter, we’ve further analysed the contribution of NTFRPs to the total energy intake, using the quantitative data from the previous chapters.

7.2 The contribution of NTFRPs to the diet

Method In the first part of the interview, the respondents were asked to quantify their farm yield in the previous season (the ‘08/’09 season, which was a good season) for field crops, garden crops, fruits and animal products. The total yield was further classified into consumed and sold produce and into several smaller categories such as ‘feed’, ‘labour payment’ and ‘storage’ (see: Table 7.1). For consumed produce, the period of consumption was determined (e.g. October-December) and for sold produce, the returns were quantified in US dollars. Additionally, the respondents were asked to quantify purchased or donated foodstuff per month. To triangulate these answers, the respondents were asked to describe what they had eaten in the past week, including main meals and snacks. Since the last week is close to memory, we assumed that the respondents would recall relatively well what they had eaten during that period. In the next section of the questionnaire, the respondents were asked to quantify their farm yield in bad years. They tended to quickly answer that they produced ‘nothing’ in bad years, but further questioning usually gave some numbers. However, there were some controversies; some farmers, for example, claimed to have harvested nothing from their gardens, while others indicated to have had an abundant harvest of garden crops. This may have to do with differences in location or water availability. Using a number of conversion factors and assumptions (see: Appendix A), the values given by the farmers were translated to kilograms and the yield per product per farmer (yi) was determined. The total yields (Y) of product p per resource group r were calculated by adding together the yields of all the individual households in the resource group (see: equation 7.1). The total yields per product p over the whole sample ( Yt ) were calculated by summing up the yields over all resource groups (see: equation 7.2). It must be noted that these yields are an estimate because in Ushe, the yield of garden crops was quantified less rigorously and therefore it is likely that some produce was left out. The yields are shown in Table 7.2.

61 Eq. 7.1 : Yield of a farm product per resource group per year.

n r = r Y , yp ∑ yi ,, yhp , h=1 where Y is the total yield of product p per resource group r in year y (in kg/year), and yi is the yield of product p per household h (h = 1, 2, ..., n) (in kg/year).

Eq. 7.2 : Yield of a farm product for all households.

n = r Yt , yp ∑Y , yp , r =1

where Yt is the overall yield of farm product p in year y over all households from all resource groups (r = 1, 2, 3) (in kg/year).

To the yield per product per farm ( yi), the purchased (pu ) and donated (do ) quantities of the product were added and the sold ( so ), stored (st ), paid (for labour, la ) or otherwise utilised produce (ot ) was subtracted, resulting in yearly amounts available for consumption per product per household ( C) (see: equation 7.3). These data were, if possible, triangulated with the data on the food consumption of the previous week. Consumption of product p in month m ( MC m) was calculated by dividing the yearly amounts available for consumption over the number of months that the product was consumed ( n, for example from May to October: n = 6) and then multiplying this number by a monthly consumption index i ( i = 1 if product p is consumed in month m, otherwise i = 0) (see: equation 7.4).

Eq. 7.3 : Yearly amounts available for consumption per product per household.

= + + − − − − C ,, yhp yi ,, yhp pu ,, yhp do ,, yhp so ., yhp st ,, yhp la ,, yhp ot ,, yhp , where C is the amount of product p per year per household available for consumption for household h in year y (in kg/year), yi is the yield, pu is the purchased amount, so is the sold amount, st is the stored amount, la is the amount used for paying labourers, and ot is the amount spent in other ways.

Eq. 7.4 : Monthly amounts available for consumption per product per household.

C = ,, yhp MC ,,, ymhp *i ,,, ymhp , n ,, yhp

where MC is the amount of product p per month per household available for consumption for household h in month m in year y (kg/month), n is the number of months per year that the product is consumed and i is the monthly consumption index. If the product is consumed in month m, then i = 1, otherwise i = 0.

62

The farm products were divided into five categories (based on the FAO World Programme for the Census of Agriculture 2010) with a few adaptations. Category 1 is the cereals (including e.g. maize and finger millet) with the addition of starchy root crops (such as sweet potato, Irish potato and cassava), category 2 (FAO group 7) is the legumes (including cowpea and soybean), category 3 (FAO group 2) is the vegetables and melons (including leaf vegetables, onions, tomatoes, pumpkins and carrots), and category 4 (FAO group 3) were fruits and nuts (including mangoes, sugarcane and avocado). Additionally, there is the category of animal products (milk, meat and eggs) and herbs/spices/flavouring (including salt and sugar). The herbs/spices/flavouring category was excluded from the further analysis because of its very limited contribution to the overall energy intake. Thus, there were five categories in total: c1. Starch crops c2. Legumes c3. Vegetables c4. Fruits c5. Animal products

The energy contribution of product p in month m per person per household ( EP p) was calculated by multiplying the amounts available for consumption in kg per product per month per household ( CM ) with the energy content of the product ( u) (FAO, 1968) and dividing this over the number of household members (excluding infants) or people eating along ( n) (see: equation 7.5). The energy contribution per category (EC c) was calculated by summing up the energy contributions of all the products in a category (see: equation 7.6). Then, the average energy contribution per season s (ES s) was calculated by summing the energy contributions of all months in the season and dividing over the number of months (see: equation 7.7). Finally, this number was averaged per resource group, resulting in E, which is the average energy availability per person per month for each category, each season and each resource group (see: equation 7.8). The results, in the form of E values, are shown in Table 7.3.

Eq. 7.5: Monthly contribution of a product to the energy availability per individual.

CM *u = ,,, ymhp p EP ,,, ymhp , nh

where EP is the energy from product p, available per member of household h in month m in year y (MJ/person/month), u is the energy content of the product (MJ/kg) (FAO, 1968) and n is the average number of people eating along per meal. Usually, n is the number of household members, excluding infants.

63 Eq. 7.6 : Monthly contribution of a product category to the energy availability per individual.

n = c EC ,,, ymhc ∑ EP ,,, ymhp , p=1

where EC is the energy from product category c, available per member of household h in month m in year y (MJ/person/month).

Eq. 7.7: Monthly contribution of a product category to the energy availability per individual per season.

n s ∑ EC ,,, ymhc = m=1 ES ,,, yshc , as

where ES is the average energy from product category c per month, available per member of household h in season s in year y (MJ/person/month), EC is the energy consumption in month m (m (within season s) = 1, 2, 3) and a is the number of months in season s. The seasons are early dry (ED, June to August), late dry (LD, September to November), early rain (ER, December to February) and late rain (LR, March to May).

Eq. 7.8: Monthly contribution of a product category to the energy availability per individual per resource group per season.

n r ∑ ES ,,, yshc = h=1 E ,,, ysrc , br

where E is the average energy from product category c per month, consumed by an individual from resource group r in season s in year y (MJ/person/month), ES is the energy consumption per person for household h (h (within resource group r) = 1, 2, ..., n) and b is the number of households in resource group r.

To analyse the contribution of edible NTFRPs to the energy availability, the edible products from Table 6.1 were divided over the same categories as the cultivated foods and their energy contribution per resource group per season was calculated according to equation 7.4 to 7.8. The calculated values were included in Table 7.3.

64

Table 7.1: Uses of the produced maize in Dendenyore Ward Maize use by farmers in Dendenyore, good year. Household D1 D7 D8 D10 D5 D9 D2 D3 D4 D6 D11 Resource group 1 1 1 1 2 2 3 3 3 3 3 Total quantity produced (kg) 1800 1700 750 1000 1050 650 800 900 1270 1600 300 Consumed (kg) 860 400 550 500 800 300 600 675 1220 1300 300 Consumption period (# months) 12 12 12 13 10 10 12 12 9 14 8 Sold/bartered (kg) 1000 200 250 Stored (kg) 500 Given away (kg) 250 Feed (kg) 240 150 200 100 50 17 75 50 50 Seed (kg) Labour payment (kg) 200 150 250 150 250 150

Table 7.2: The added total yield per year per RG per product (kg fresh weight/year) in good and bad years GOOD BAD Product name Production (kgFW/year) Production (kgFW/year) RG 1 RG 2 RG 3 TOTAL RG 1 RG 2 RG 3 TOTAL Starch crops Maize 16524 5575 6470 28569 1030 840 617 2486 Sweet potato 1021 106 136 1263 0 615 0 615 Finger millet 739 144 99 981 161 18 0 179 Sunflower 380 240 70 690 0 0 40 40 Irish potato 35 18 354 407 0 0 0 0 Sorghum 181 72 90 343 0 0 0 0 Wheat 0 0 323 323 0 0 108 108 Cassava 0 18 0 18 0 18 0 18 Sunhemp 0 0 3 3 0 0 0 0 Pearl millet 0 0 0 0 0 18 0 18 TOTAL 18881 6173 7544 32598 1191 1509 764 3465 Pulses Cowpea 1105 239 157 1501 0 18 0 18 Groundnut 831 372 244 1447 119 176 26 322 Soybean 896 50 166 1112 50 0 0 50 Roundnut 435 242 216 892 21 17 35 73 Beans 92 78 101 272 0 37 28 64 Peas 18 37 6 61 0 0 0 0 Sugarbeans 10 0 0 10 0 0 0 0 TOTAL 3388 1018 890 5295 190 249 88 527

65 Table 7.2, continued GOOD BAD Product name Production (kgFW/year) Production (kgFW/year) RG 1 RG 2 RG 3 TOTAL RG 1 RG 2 RG 3 TOTAL Vegetables Leaf vegetables 1535 585 1959 4079 548 183 1587 2317 Tomatoes 647 185 435 1268 98 120 16 234 Sweet cabbage 759 0 0 759 0 0 0 0 Onions 375 41 34 450 0 30 0 30 Pumpkin 200 100 0 300 0 100 0 100 Butternut 30 0 100 130 0 0 0 0 Carrots 11 22 0 33 0 0 0 0 Spring onions 15 10 5 30 0 0 0 0 Cucumber 24 0 0 24 0 0 0 0 Okra 20 0 0 20 0 0 0 0 TOTAL 3616 943 2533 7092 645 432 1603 2681 Fruits Sugarcane 1561 125 0 1686 0 18 0 18 Mango 1053 492 0 1545 720 480 0 1200 Guava 516 0 24 540 0 0 12 12 Banana 68 250 38 355 0 150 20 170 Peach 54 36 216 306 0 24 0 24 Avocado 204 0 0 204 0 0 0 0 Pawpaw 164 20 0 184 0 0 0 0 Orange 120 44 0 163 0 0 0 0 Lemon 65 0 0 65 0 0 0 0 Naartjie 41 0 0 41 0 0 0 0 Grape 20 0 12 32 0 0 0 0 Apple 2 0 0 2 0 0 0 0 TOTAL 3866 967 290 5122 720 672 32 1424 Animal products Milk (litres) 780 0 0 780 0 0 0 0 Chicken meat 164 135 50 349 39 105 51 195 Broilers 300 0 0 300 0 0 0 0 Goat meat 135 90 0 225 30 210 0 240 Cattle meat 216 0 0 216 0 144 0 144 Chicken eggs 63 50 42 156 520 0 456 976 Turkey meat 48 0 0 48 48 0 0 48 Guineafowl meat 0 8 0 8 0 0 0 0 Fowl eggs 0 5 0 5 0 0 0 0 TOTAL 1706 288 92 2086 637 459 507 1603

At times of food insufficiency, notably after the ‘07/’08 drought and after the ’91/’92 drought, donors were providing farmers with food. The donors were mainly active from October to May in the year following upon the drought. Food was provided in the form of maize or maize meal, peas or beans, porridge and cooking oil. Most farmers indicated to have received food-aid, but some claim that they did not. Indeed, donors did not give food to all farmers but made a selection based on a certain classification. There was large variation in the amount of food that was available to the farmers. Some farmers had the fortune of having family members that could provide them with additional food. Others would trade a goat for two buckets of maize (own observation). All these additional sources of food were quantified so that

66 they could be included in the total diet. For animal products, it was assumed that the same amount of eggs and poultry meat was consumed in good and bad years (unless specifically stated otherwise) but that there was no cattle or goat meat consumption in bad years (unless specifically stated otherwise) because those animals were either preserved or used for barter trade, but not consumed as such.

Results Table 7.1 shows the use of a maize harvest in a good year in Dendenyore. Out of eleven farmers, three sold part of their maize to generate cash income. Nine used some of the harvest to feed the chicken, turkeys or guinea fowls. Six, including three from RG 1, two from RG 2 and one from RG 3, used part of the maize to hire labour during the cropping season. Only one farmer indicated to store a major part of her yield, while four others, including two from RG 2 and two from RG 3 indicated that they did not have enough maize to eat the whole year.

Table 7.2 shows the total farm production per resource group and for the entire sample. It also shows the entire range of products that farmers produced, and the division over the different categories. Our sample of 25 farmers produced over 28000 kg of maize per year. They also produced over 4000 kg of leaf vegetables (kale and rape), over 1600 kg of sugarcane and over 1500 kg of cowpea and mangoes. In bad years, the production dropped dramatically. Over 90% less maize was produced, as well as 50% less leaf vegetables and 90% less pulses. The enormous decrease in production affected each of the resource groups, and even though the RG 1 farmers still produced most of the maize, for the other products there was little difference left between the groups. In good years, the results suggest that RG 1 farmers produced more animal products, more fruits, more pulses and more starch crops than RG 2 and RG 3 farmers.

Table 7.3 shows the energy availability in good years (a) and in bad years (b) per product category per resource group per season in MJ/person/month. In good years, RG 1 farmers had the highest energy availability, which was almost 50% higher than the intake of RG 2 and RG 3 farmers. RG 1 farmers produced more pulses and significantly more maize than RG 2 and RG 3 farmers, which was the main reason for their higher availability. RG 2 farmers had a slightly higher availability of energy from pulses, cultivated fruits and animal products but a lower availability of starch crops than RG 3 farmers, resulting in an overall energy availability that was marginally lower for RG 2 than for RG 3 farmers. Altogether, RG 1 farmers had the richest diet, with a monthly energy availability of approximately 470 MJ per person and a relatively large share of pulses, vegetables, wild and exotic fruits and meat. RG 3 farmers, on the other hand, had an average availability of only 331 MJ per person, with a much lower contribution of exotic fruits, pulses and meat (see: Figure 7.1). For all resource groups, the relative contribution of cultivated fruits, vegetables and animal products to the total energy availability was low, but of course these types of food supplied other essential elements to the diet such as proteins and vitamins. Only banana, avocado and the wild fruits (notably Uapaca kirkiana , Parinari curatellifolia and Strychnos spinosa ) had a relatively high energy content and thus contributed significantly to the energy supply. The seasonality of the total energy availability was different for each of the resource groups. In general, the deviations per season were no more than 13% above or below the average, and the availability was above the minimum of 229 MJ/person/month for

67 all resource groups in all seasons. RG 1 farmers had the highest energy availability in the early rain season, mainly due to the high contribution of exotic and wild fruits in this period. RG 2 farmers had the highest energy availability in the early dry season, because of a high contribution of starch crops, pulses and wild fruits. RG 3 farmers had their peak availability in the late dry season, because of a high starch crop contribution and a relatively high wild fruit consumption. None of the resource groups had the lowest availability in the late dry season or in the early rain season, because the ripening of wild and exotic fruits mainly happens in these periods. RG 1 farmers had the lowest energy availability in the early dry season because the consumption of wild fruits and starch crops was low. Both RG 2 and RG 3 had the lowest energy availability in the late rain season. In this period, the starch crops reserves were running out and the consumption of wild fruits and wild animals/insects was relatively low.

In bad years, crop production was greatly reduced (see: Table 7.2), which also resulted in a large reduction in energy availability (see: Table 7.3b). For the RG 1 farmers the difference was largest, with an availability reduction of over 40%. This reduction was caused mainly by the lower production of starch crops, pulses and exotic fruit. The RG 3 farmers on the other hand had an energy availability that was only marginally reduced, from 331 to 326 MJ/person/month. RG 3 farmers had a relatively low contribution of starch crops and pulses to the total energy availability in bad years, but a relatively high contribution of vegetables. For RG 2 farmers, the contribution of starch crops and pulses decreased but the contribution of animal products increased. For all resource groups, wild fruits became a major contributor to energy availability, predominantly in the late dry season (see: Figure 7.1). Especially RG 3 farmers greatly increased their wild fruit consumption, pushing their total energy availability up to 478 MJ/person/month in the late dry season, when most indigenous fruits were ripe. RG 2 and RG 1 farmers consumed much less wild fruits, resulting in an overall seasonal and yearly energy availability that was higher for RG 3 farmers (326 MJ/person/month) than for RG 1 and RG 2 farmers (264 and 286 MJ/person/month, respectively).

Figure 7.1 shows the composition of the diet of the different resource groups in good and in bad years. In good years, the poor farmers mainly got their energy from starch crops, which was mostly maize. The wealthier farmers consumed relatively more pulses, meat and exotic fruits than the poor farmers, but the relative consumption of vegetables and wild fruits was similar for all wealth groups. In bad years, RG 3 farmers consumed relatively even less pulses and exotic fruits compared to the other resource groups. However, their consumption of vegetables, wild starch crops and wild animals/insects increased in bad years, as well as the consumption of wild fruits. The dependence of RG 3 farmers on maize decreased in bad years, as wild fruits started to contribute a large share of the energy. The relative dependence of the wealthier farmers on maize also decreased in bad years, but to a lesser extent. Whereas RG 3 farmers obtained 51% of their energy from maize and 39% from wild fruits in bad years, RG 1 and RG 2 farmers obtained between 61 and 65% of their energy from maize and between 19 and 21% from wild fruits.

68 Table 7.3: Energy availability (MJ/person/month) in a) good years and b) bad years per resource group, divided over the different product types. ED = early dry, LD = late dry, ER = early rain, LR = late rain. a) Energy availability (MJ/person/month), GOOD YEAR Product Type Season RG 1 RG 2 RG 3 AVE STD AVE STD AVE STD Starch and energy crops Cultivated ED 319 113 269 65 264 138 LD 318 118 268 65 282 196 ER 330 136 247 80 266 209 LR 346 137 247 85 234 189 Wild ED 0 0 0 0 0.2 0.4 LD 0 0 0 0 0.0 0.0 ER 0 0 0 0 0.0 0.0 LR 0 0 0 0 0.1 0.3 Pulses Cultivated ED 67 49 48 26 42 25 LD 65 50 41 21 34 26 ER 60 55 19 10 17 21 LR 65 50 36 20 32 27 Vegetables Cultivated ED 7 11 5 2 4 2 LD 9 12 4 2 7 7 ER 8 11 3 2 4 3 LR 5 9 4 1 3 3 Wild ED 0.2 0.5 0.0 0.0 0.0 0.0 LD 0.9 2.0 0.0 0.0 0.0 0.1 ER 2.3 4.1 1.6 1.7 0.7 1.1 LR 0.2 0.5 0.0 0.0 0.2 0.5 Fruits Cultivated ED 25 38 5 10 0.3 0.4 LD 20 36 4 8 0.3 0.4 ER 38 53 16 20 8.2 15 LR 26 39 10 14 15 35 Wild ED 0.1 0.2 46 91 1 3 LD 66 128 19 28 39 41 ER 43 51 0 0 52 64 LR 0 0 15 30 0 0 Animals/insects Domesticated ED 12 10 6 4 3 3 LD 12 10 6 4 3 3 ER 12 9 10 12 4 6 LR 12 9 6 4 4 6 Wild ED 1 2 0 0 1 1 LD 5 8 11 19 1 1 ER 6 7 32 56 2 3 LR 1 2 1 1 1 1 TOTAL ED 431 379 315 LD 494 352 366 ER 499 329 353 LR 456 319 289 ALL YEAR 470 345 331

69

b) Energy availability (MJ/person/month), BAD YEAR Product Type Season RG 1 RG 2 RG 3 AVE STD AVE STD AVE STD Starch and energy crops Cultivated ED 161 100 153 86 156 152 LD 176 80 170 69 165 147 ER 185 68 179 56 183 137 LR 169 89 202 137 164 147 Wild ED 0 0 0 0 28 58 LD 0 0 0 0 0.0 0.0 ER 0 0 0 0 0.0 0.0 LR 0 0 0 0 3.2 7.2 Pulses Cultivated ED 18 20 20 17 5.9 8 LD 20 17 23 13 11 10 ER 22 16 23 11 13 10 LR 19 18 21 15 7.5 8 Vegetables Cultivated ED 2 3 3 3 8 13 LD 4 5 2 2 8 13 ER 4 7 2 3 8 13 LR 2 2 2 2 8 13 Wild ED 0.4 1.1 0.0 0.0 0.0 0.0 LD 0.4 1.1 0.0 0.0 0.0 0.0 ER 0.4 1.1 1.0 1.3 2.1 4.8 LR 0.4 1.1 0.0 0.0 0.0 0.0 Fruits Cultivated ED 0 0 3 7 0 1 LD 0 0 4 6 0 1 ER 17 41 20 33 0 1 LR 0 0 3 6 1 1 Wild ED 0 0 61 122 50 74 LD 153 189 129 143 288 255 ER 67 58 0 0 166 120 LR 0 0 23 29 0 0 Animals/insects Domesticated ED 8 7 14 21 1 1 LD 8 7 14 22 1 1 ER 8 7 15 21 4 6 LR 8 7 14 21 4 6 Wild ED 1 2 0 0 4 5 LD 1 2 9 19 4 5 ER 2 3 28 56 4 6 LR 1 2 4 8 4 5 TOTAL ED 190 256 253 LD 361 351 478 ER 305 268 380 LR 198 270 192 ALL YEAR 264 286 326

70

Figure 7.1: Energy provision by different crop and product types in a) good years and b) bad years, per resource group. C = cultivated/domesticated, W = wild.

7.3 Discussion

Using the yields of the different farm products, the allocation of these products (own use, sales, labour payment, etc.) and the energy content per product, we calculated how much energy was available for consumption per household member per year in good years and in bad years. According to the FAO, the minimum energy intake of an average Zimbabwean is 1800 kcal/day, or 229 MJ/month. People who take in less than this minimum per day are considered ‘hungry’. Our data indicate that in a good year all farmers had, on average, an energy availability well above the minimum (see: Table 7.3). In the Netherlands, the advised maximum intake per day for an adult male is 2500 kcal, which comes down to 318 MJ/month. Our data are reasonably well within this range, but the energy availability for RG 1 farmers is quite high. It is unlikely that farmers actually consumed all the available energy (470 MJ per month) in good years, and there are several reasons why the available energy may not all be consumed. First of all, we did not always include visitors and members of the

71 extended family in our estimation of the number of people that would eat from the household production. It may very well be that especially some of the wealthier families also provided for others than the direct family members. If the food is shared with more, then naturally the consumption per person will go down. Additionally, it is likely that the storage of grains and pulses has been underestimated. In our sample, only two farmers indicated that they had stored significant amounts of their harvest, but probably many more had surpluses that were retained. It is likely that we formulated the question regarding storage incorrectly. If we assume that the actual retention of especially maize, small grains and pulses is higher than our data show, and that mainly the RG 1 farmers had surpluses to store, then we must assume that the energy consumption of especially the RG 1 farmers in a good year was lower than the energy availability that we calculated. Also, the underestimation of storage may have led to underestimation of the energy availability in bad years. One of the farmers who indicated to have stored significant amounts of food also indicated that he had some stored maize and beans which he used to get him and his family through the bad year of ‘07/’08. It is likely that other farmers, especially from RG 1, had some surpluses at the start of the bad year, too, and therefore the energy availability in especially the beginning of the bad years was probably higher than our data show. This may also be the reason why RG 1 farmers collected less wild fruits than RG 3 farmers.

On average over the entire bad year, farmers of all resource groups had an energy availability above the ‘hunger’ line. For the RG 2 farmers, the energy availability was above the hunger line in all seasons, but for the RG 3 farmers the energy availability dropped below the critical line of 229 MJ per person per month in the late rain season and for RG 1 farmers it dropped below the critical line in the early dry and in the late rain season, when wild foods were scarcely available and food donors were not active. However, it is likely that RG 1 farmers used surpluses from the previous year to overcome the deficits. To get a picture of the total energy availability, we included the donated food in the diet. It would have been informative to also calculate the total energy availability excluding the donated food, because not all farmers receive food-aid and, more importantly, because food-aid is an emergency mechanism. Farmers need to be able to provide for their families without it, in order to be truly food secure.

As Table 7.3b shows, the consumption of some NTFRPs, especially wild fruits, greatly increased in bad years. Approximately 22% of the available energy for RG 1 and RG 2 farmers and as much as 42% of the available energy for RG 3 farmers was supplied by NTFRPs in bad years. These results suggest that wild foods make a relatively large contribution to the diet of farmers in bad years, and they support our hypothesis that wild foods are essential to prevent food insecurity or hunger. Without this contribution, almost all farmers would be food insecure during the entire dry season after a bad year. Especially the poor farmers (RG 3) appear to benefit from the wild fruits, which supports our hypothesis that poor farmers are more reliant on NTFRPs than rich farmers. However, the differences were not significant.

72 Chapter 8: Livestock management and feed

8.1 Introduction

The communal woodlands and rangelands are not only important as a source of NTFRPs for use and consumption by people, but also as a source of livestock feed (Scoones, 1995; Campbell et al. , 2000; Rubanza et al. , 2004; Kepe, 2007). Crop failure in bad years affects not only the people, but the livestock as well. We’ve assessed the importance of the NTFRPs ‘livestock feed’, collected by farmers (mainly branches, pods and grass) and ‘livestock grazing’ in good and bad years.

8.2 Livestock management in good and bad years

Methods From the total sample, six farmers in Ushe and six farmers in Dendenyore were further interviewed about livestock feed and grazing. The farmers were asked to indicate the different types of feed that the livestock was provided with and the herding strategies during different seasons and in different years. Also, farmers were asked to indicate the grazing areas on the map. Results are gathered in Table 8.1 and the indicated grazing areas are shown on Figure 8.1 and 8.2.

Results In Ushe, there were few areas of open grassland. Four out of the six farmers in Ushe indicated that their cattle grazed mainly in woodland or bushland areas with grassy undergrowth (see: Figure 8.1). Figure 8.1 shows that generally the farmers in Ushe chose the nearest uncropped area as a grazing site. Farmer U7 and U8 herded their animals towards a particular spot at the side of the large mountain close to their homestead. The majority of the farmers in Ushe did not change the herding strategy and grazing area in bad years (see: Table 8.1b and Figure 8.1). According to the farmers, the grazing areas provided sufficient feed also in bad years. However, one farmer in Ushe (U7) indicated that the free-grazing cattle stayed close to the homestead in good years but strayed all the way down to the river in bad years. In Dendenyore there were several grassland areas including some improved sweetveld for cattle grazing (own observation). Most cattle-, donkey- and goat-owning farmers in Dendenyore chose to graze their livestock in areas of veld, which were not necessarily the nearest uncropped areas nearest to the homestead (see: Figure 8.2). In bad years, three out of six farmers in Dendenyore went to a different grazing area, which provided more feed (see: Table 8.1b and Figure 8.2). It is likely that the differences in livestock management between Ushe and Dendenyore were caused by the difference in resource base. Figure 8.1 and 8.2 clearly show that Dendenyore has more areas of woodland and bushland than Ushe, and also more areas of veld. Therefore, farmers from Dendenyore were more free to select an optimal area for grazing their livestock, whereas farmers from Ushe had less room for choice.

73 Table 8.1: Livestock grazing and feed in a) good years and b) bad years. All farmers interviewed owned at least two cattle, except U10 who owned no cattle but two donkeys and D9 who owned only one donkey. Grazing area = the area where the cattle/donkeys are taken when they are being herded. a) GOOD YEAR Farmer Grazing area Herding period Additional feed 1 Additional feed 2 U1 Woodland in village Groundnut leaves Crop residues U2 Woodland in village MidNov-May Stover with salt Piliostigma thonningii pods U7 Around the mountain Oct-Jul Maize stover U8 South side of mountain MidNov-May Maize stover U9 Around the dam Dec-Jun Maize stover U10 Next to the stream Dec-Apr Green maize Maize stover D1 Bushland near Nyamidzi river Nov-May Pumpkins 0.25 cart Maize stover D5 Sourveld & around mountain Oct-May Crushed maize Salt D7 Veld nearby and next to dam Nov-Jun Stover, crushed maize Hay D8 Veld next to dust road Nov-May Maize stover D9 Next to Jekwa river Nov-May Maize stover Cowpea & beans stover D10 Sourveld around river Oct-May Maize & pulses Thatch grass

b) BAD YEAR Farmer Grazing area Herding period Additional feed 1 Additional feed 2 U1 Woodland in village Nov-Jun None U2 Woodland in village MidNov-May Piliostigma thonningii pods U7 Around the mountain Oct-Jul None U8 South side of mountain MidNov-May None U9 Around the dam Dec-Jun None U10 Next to the stream Dec-Apr None D1 Everywhere, long distances None D5 Sourveld & around mountain Oct-May None D7 Veld nearby and next to dam Nov-Jun Branches of trees D8 Along Nyamemba river None D9 Next to Jekwa river Nov-May None D10 Woodland beyond borders Summer Thatch grass

The cattle in Ushe and Dendenyore were herded only during the rain season, to keep the animals from destroying the crops. In the dry season, the livestock was free to graze anywhere, apart from the gardens which were fenced to keep out the cattle. Herded cattle is known to have a higher intake of energy than free-grazing cattle (Turner et al. , 2005) but in good years, cattle were able to find sufficient feed also when unherded. However, in bad years, feed availability became a serious problem. Nevertheless, none of the farmers extended the period of herded grazing in bad years. This is probably a management choice based on labour availability; herding cattle is a very labour-intensive activity and this labour can be saved by letting the cattle graze freely (Wolmer and Scoones, 2002).

74

Figure 8.1: Bushland in Ushe Ward.

8.3 Energy provision for livestock from crop residues and grazing

Methods Eleven out of twelve farmers indicated that they collected maize stover after harvesting, to feed to the cattle as supplementary food at the end of the dry season. Maize stover was the main source of energy in this period, when crop residues had all been grazed from the fields and the grass was no longer palatable. The amount of energy that was provided to the cattle in the form of maize stover per household ( HE ) was calculated by multiplying the maize yield per household ( Y) with the average maize harvest index ( HI ), the average stover dry matter content ( DMC ) and the average stover energy content ( MEC ) (see: equation 8.1) (Lopez et al. (2005)). The energy contribution per head of livestock (only cattle and donkeys) for household h (LE ) was calculated by dividing the energy per household over the number of livestock in that household ( n) (see: equation 8.2).

Eq. 8.1 : Energy availability for feed (in the form of maize stover) per household.

= HE maize ,, yh Ymaize ,, yh * HI maize * DMC maize * MEC maize ,

where HE is the household energy availability from product maize in household h in year y (MJ/household/year), Y is the maize yield, HI is the harvest index (0.50 for maize), DMC is the dry matter content (0.80 for maize stover) and MEC is the metabolisable energy content (6.3 MJ/kg for maize stover) (Lopez et al. , 2005).

75 Eq. 8.2: Energy from maize stover per livestock head per household.

HE = maize ,, yh LE maize ,, yh , nh

where LE is the energy from product maize per head of livestock, for household h in year y, and n is the number of livestock for household h.

An average oxen of 250 kg has a maintenance energy requirement of 5800 MJ/year (Pearson, 1996). In good years, the amount of energy provided to livestock in the form of stover by the different households ranged from 966 to 3780 MJ per year, with an average of 1972 MJ/year. Most of the remaining energy was provided by the grass and browse from the rangelands and woodlands of the community, as well as by crop residues that remained on the fields after harvesting. In bad years, maize growth was significantly decreased (p < 0.001) and little or no stover was available to feed to the cattle. Thus, the animals became almost solely dependent on graze and browse from the community lands (see: Figure 8.1). Two farmers (U2 and D7) indicated that they collected Piliostigma thonningii pods for feeding the cattle, both in good and in bad years. P. thonningii is a good source of protein, with a crude protein content of 10.1% of DM (FAO, AFRIS (no date)). Farmer D7 also collected tree branches to feed the cattle (the type of tree was not specified) in bad years only. Another farmer (D10) collected thatch grass to feed to the cattle at the end of the dry season, both in good and in bad years. The remaining farmers left their cattle to be solely dependent on graze and browse as source of feed during the late dry season in bad years, supplemented with the little stover that they had (see: Figure 8.1). This put extra pressure on the vegetation, and several farmers indicated that the feed availability was insufficient and that their livestock had died during the bad year.

Figure 8.1: Energy provision for cattle supplied by maize stover (black) and free range (grey) in good years and bad years.

76 8.4 Discussion

Our data suggest that if areas of veld are available, farmers will preferably graze their cattle in those areas. In bad years, some farmers in Dendenyore chose to graze their cattle in different areas than in good years, but farmers in Ushe did not, which may reflect either the sufficient availability of feed in the selected grazing areas in Ushe or the lack of options that farmers in Ushe have. In good years, maize stover was used to feed the cattle at the end of the dry season. In bad years, little stover was available and the livestock energy intake became more dependent on resources from the common lands. However, farmers did not increase the period of herded grazing to increase the energy intake.

Figure 8.1: Land use areas in Ushe with indicated grazing areas, on a Google Earth screenshot. Light green= bushland; dark green= woodland; yellow= veld; red= ward border; blue= river/stream; white= main road; yellow pins= farms & special buildings. Source: Google Earth TM , with modifications.

77

Figure 8.2: Land use areas in Dendenyore with indicated grazing areas, on a Google Earth screenshot. Light green= bushland; dark green= woodland; yellow= veld; red= ward border; blue= river/stream; white= main road; yellow pins= farms & special buildings. Source: Google Earth TM , with modifications.

78 Chapter 9: Discussion and conclusions

9.1 Discussion

Farmers collected a range of products from the communal lands The farmers that we interviewed together were able to name over 130 NTFRP species that were collected in their community. However, not all of these NTFRPs were collected regularly or considered as important. We asked the respondents to select five NTFRPs that were very important to them in good years and five that were very important in bad years of crop failure. Most farmers found it difficult to make this selection. They took a long time to choose, they started picking on top of the product list and forgot the products at the bottom of the list, they only picked species from one product type or they forgot to select important NTFRPs such as firewood. Luckily the next exercise, to place the five NTFRPs in an order from most important to least important, was more easily done, because the selection was already made. Out of the list of approximately 55 NTFRPs that the farmers considered as ‘important’, we selected 17 products for further analysis. Fifteen of those were the highest ranking products in the list. We limited ourselves to these fifteen because we wanted to have a wide range of products to analyse, but for the next products in rank, too little quantitative data per product was available to draw any statistically sound conclusions. However, there were no products of the categories ‘vegetables’ and ‘roots/tubers’ among the 15 highest ranking products. Therefore, we decided to add Corchorus oditorius and Coleus esculentus to the selection, which were the highest ranking products in the ‘vegetables’ and the ‘roots/tubers’ category, respectively. Thus, we came to a selection of 17 key NTFRPs (four non-food products, five fruit species, one mushroom species, one vegetable species, two types of animals, two types of insects, one tuber and one herb) on which we did further analysis.

Two important NTFRPs were the fruits U. kirkiana and P. curatellifolia The purpose of our project was to provide scientific support for the hypothesis that NTFRPs can serve to prevent or reduce food insecurity at times of crop failure due to poor rainfall. Our data show that wild fruits ( P. curatellifolia and U. kirkiana ) were consumed and/or collected at increased quantities in bad years. In such years, these fruits appeared to contribute greatly to the overall energy availability of farmers, especially in the late dry season. Their contribution is what made the difference between food insufficiency and food sufficiency. Mithöfer and Waibel (2003) studied the use of wild fruits in Zimbabwe, and they came to the conclusion that especially P. curatellifolia , and to a lesser extent also S. spinosa and U. kirkiana , were consumed as a meal, rather than as a snack, at times of poor harvest. Our data agree with these findings, and farmers in our sample indicated to use P. curatellifolia not only to eat raw, but also to make a drink, to bake bread and to prepare porridge in bad years. Besides the relatively high energy content of U. kirkiana and P. curatellifolia , there are other reasons why these fruits are very suitable as energy source in years of drought. Respondents indicated that both P. curatellifolia and U. kirkiana flower in response to stress, which results in very abundant fruit yields in the year following upon a drought, unlike exotic fruits which tend to yield very poorly at such times. Also, both P. curatellifolia and U. kirkiana fruit from the end of the dry season to the beginning of the rain season (Mithofer, 2004). This is a period in which little other food is available because stocks are depleted but the new harvest is not yet ripe. Thus, the abundant P. curatellifolia and U.

79 kirkiana fruits come at the right time to fill the gap in energy provision that is left by failed maize harvests, poorly yielding exotic fruits and a generally unproductive time of the year.

There is an unclear relationship between NTFRP use and household wealth We divided the respondents in our sample into three different resource groups: RG 1 (wealthy), RG 2 (medium-wealthy) and RG 3 (poor). These three groups were classified based on assets only, but after classification we found that there were other differences as well. RG 1 farmers produced significantly more maize than RG 3 farmers, which can be explained because RG 1 farmers have more land and can purchase more inputs, while they also have manure and draught power available from their cattle. Also, the household heads from RG 1 were significantly older than the heads of the RG 3 households. As an explanation, we propose that assets are collected, purchased or otherwise obtained over time, and therefore the older farmers are likely also to be the richer ones. To add to that, most RG 1 households (but none of the RG 3 households) were supported by children working in the city. In contrast, all RG 3 farmers but one had only children that were still under age. Though the average number of adults per household appeared to be higher for RG 1 than for RG 3 farmers, the total number of household members was rather similar. Despite the differences between RG 1 and RG 3 households, we did not find any significant differences between the resource groups regarding the total consumption of key NTFRPs per household. These results do not agree with the results of Cavendish (2000) who found in a study in Zimbabwe that poor households derived a larger share of their income from NTFRPs, but that the absolute use quantities of NTFRPs were higher for wealthy households. However, they are in agreement with the results of Cocks et al. (2008), who concluded that household wealth did not significantly influence NTFRP use, at all. From an economic perspective, the influence of household wealth on NTFRP use is also debated. In a study in three villages in South-Africa, Shackleton and Shackleton (2006) found that wealthy households bought NTFRPs significantly more often than poor households, but poor households were found to derive a higher annual direct-use value from wild herbs than wealthy households did. And in a recent study of Paumgarten and Shackleton (2009) the data suggested that poor households were more dependent on NTFRPs because they sold a greater variety of products and bought significantly less NTFRPs than wealthy households. All studies underwrite the conclusion of Luckert et al. (2000) that the relation between household wealth and NTFP use deserves more specific attention.

Our data show that the labour allocation to the collection of NTFRPs also did not differ between the resource groups. However, only the wealthier (RG 1) households invested significantly more labour in the collection of an NTFRP (namely U. kirkiana ) at times of drought. This significant increase shows the importance of the edible NTFRP U. kirkiana for wealthy households at times of crop failure, and it does not support the statement that poor farmers are relying more on NTFRPs than wealthier farmers. However, when looking at the diet of wealthier and poorer farmers, both in good and bad years, one can see that the contribution of NTFRPs, especially wild fruits, to the total energy availability was much higher for poor farmers than for wealthier farmers. Unfortunately our sample was too small to obtain significant results in this respect, and therefore the hypothesis that poor farmers are more dependent on NTFRPs could not be properly tested.

80 Group discussions and increased sample size could improve the reliability of our quantitative data The data in our study have been collected mainly by in-depth interviewing of a sample of farmers. Our quantitative data were provided by the respondents, based on their recall. It is likely that the numbers are somewhat inaccurate; after all, it is not easy to remember how much one has eaten in a week in the previous year. Paumgarten (2005) states that there is ‘a lack of empirical data on the real strength of the rural safety-net function of NTFPs’ because ‘households do not depend on them on a daily basis and therefore may not be able to give accurate reports on the quantities and species used’. Regarding the ‘species used’, our extensive list of collected species shows that this statement is not correct. Farmers were extremely well able to name the species of NTFRPs they collected. Especially elder farmers had experienced many events of drought and food shortage, and they were very knowledgeable regarding the species of wild food plants that were available in their community. Regarding the quantities, indeed we must assume that there is inaccuracy, but this does not make the data useless. In fact, farmers were able to answer, with great confidence, questions such as ‘how often do you consume/collect this product?’ and ‘how much do you consume/collect each time?’. Cavendish (2000) observed that in a valuation exercise, respondents’ estimates produced values with acceptable properties, which is in agreement with our own observations. When comparing the NTFRP use that we found with quantities found in other studies, there are differences as well as similarities. For example, our results of total firewood consumption are in agreement with the findings of Grundy et al. (1993) but are relatively high compared to the findings of Vermeulen et al. (1996). Campbell (1997) found that farmers in Matendeuze, Zimbabwe, consumed 47 kg of mushrooms per year but only 0.72 kg of termites (this study: 36 kg and 21 kg, respectively). Mithofer and Waibel (2003) found that in Murehwa communal area in Zimbabwe, households spent 24 hours per year on the collection of P. curatellifolia and Strychnos spinosa , together, and 96 hours per year for the collection of U. kirkiana (this study: 104 and 124 hrs/year, respectively). Shackleton et al. (2002) found an average yearly wild fruit consumption of 116 kg per household per year, an average insect consumption of 161 kg per household per year and an average wild herb/vegetable consumption of 93 kg per household per year (this study: 378, 24 and 27 kg, respectively). And Twine (2003) found a bushmeat consumption of 2.9 kg per year, on average (this study: 25 kg). When we compare the values above with the values found in our study, we find deviations both to the higher and the lower range, while some values (such as mushroom consumption and labour expense for the collection of U. kirkiana ) are highly similar. However, we do not find a structural under- or overestimation of consumption. Still, the value of our data must be critically viewed, especially because the variation within and between the answers of our respondents was very high. For P. curatellifolia , for example, the estimate of consumed amounts was at least two times more or less than the estimate of collected amounts for 11 out of 21 respondents, and at least five times more or less for three out of 21 respondents. Even though there was no significant difference between the estimates for ‘collection’ and ‘consumption’ over the entire sample for any NTFRP but firewood, the large variation must not be ignored. We have tried to prevent overestimation of the NTFRP use by taking always the lowest value, either consumed or collected, when doing further calculations. If we assume that farmers make reasonable estimates, then it is better to take this lowest value because we assume that there was little trade or sharing of NTFRPs between farmers, so whatever was not collected could not be consumed, and whatever was not

81 consumed could simply be discarded. Additionally, taking the lowest value makes sense because consumption and collection appeared to be easily overestimated. For example, respondents would say that they collected P. curatellifolia every day in bad years, and that each time they would get a whole bucket. Other respondents would state that some days, the bucket would not be filled entirely, or not even halfway, because there were no more fruits available. Without this nuance, ‘a bucket’ sounds like ‘20 litre’, but in practice, it was probably often less. Similarly, respondents stated that they ate a ‘full plate of hacha ( P. curatellifolia ) per person every day’, but it is likely that there were also days on which they ate less, because less fruits were available or because there were other types of food to eat. These nuances were hard to capture because of the way the questions were asked and of course because of time limitations. A very useful tool that could be exploited to capture more nuance is the group discussion (see: Campbell et al. , 1997), focused on a limited number of key NTFRPs. Division of the group into for example age, sex or household assets of the respondents could provide interesting data on factors that influence NTFRP use. Additionally, the variation in estimates could be smoothed out by increasing the number of respondents. This would also make it easier to find patterns in NTFRP use. By starting with a number of pilot interviews to identify key NTFRPs and possible patterns, the questionnaire could be more specifically targeted and its length could be reduced to allow for a higher number of interviews in the same amount of time. Thus, the variation in the estimates can be smoothed out, or at least be explained better. Our study, with a sample size of 25 households only, was small. Nevertheless, our results are not structurally different from results found in other (larger) studies and we were able to find some significant patterns. Therefore, we believe that our quantitative data are a good indication, even if not proof, of the role of NTFRPs as a safety net at times of food crisis.

The energy consumption was probably lower than the energy availability because storage and consumption dynamics were only partly captured Our analysis of farmers’ food budgets has resulted in values for monthly energy availability that are structurally higher than values for energy consumption from studies described in the literature. Hallund et al. (2008) found an average intake of 228 MJ per month for Malawian women, with a range from 161 to 325 MJ per month. Vorster et al. (2004) found a monthly intake of 289±115 for healthy men in South- Africa during their study of the nutritional status of HIV infected Africans. And MacIntyre found an average monthly intake of 292±8.5 MJ for a sample of men and 240±5.5 MJ for a sample of women in rural South Africa. However, Ferro-Luzzi et al. (2002) found values for energy availability in rural Ethiopia that were comparable to ours, with an interquartile range of 135 to 480 MJ/person/month. Our average values for energy availability fall within this range, but especially for the RG 1 households in good years and for the RG 3 households in the late dry season of the bad year, it is not likely that all available energy (470 and 478 MJ/person/month, respectively) was actually consumed. In Chapter 7, we indicated that the energy consumption of the RG 1 farmers was probably lower than the energy availability because the storage of surpluses was underestimated. As Ncube et al. (2009) noted, most farmers retain their surpluses and sell them only if the next season proves to be good. We did not find this in our data, but nevertheless we consider it very likely that especially the RG 1 farmers in our sample had surpluses to store. Therefore we assume that the actual energy intake of the RG 1 farmers was well below the energy availability. Similarly, we assume that the energy availability of 478 MJ/person/month for RG 3 farmers in

82 the late dry season of a bad year is not equal to the actual consumption. In our calculations of energy availability, the energy contribution from maize in bad years was equally divided over all months, unless farmers specifically stated that the maize was finished after a certain period of time. However, it is not likely that the maize consumption was evenly spread over the entire year. Instead, we find it more likely that farmers saved maize if substitutes are available, and ran out of maize if there is not enough, rather than eating just very small amounts during the entire year. Therefore, we assume that the consumption of maize by RG 3 farmers was reduced during the late dry season, when indigenous fruits were available, and perhaps increased during other seasons when there were less wild foods, until the point where all maize was finished. Unfortunately, our methods of data collection were not suitable for getting a full picture of the actual consumption of food products throughout the year. Farmers’ diaries and perhaps group discussions could serve to provide better insights.

Woodland and rangeland areas are essential for household welfare Areas of woodland or veld provided the households in our sample with a variety of products, but many of these areas (excluding only the steep, rocky mountain slopes) could be used for agricultural production, as well. According to Mr. Ushe, the headman of Ushe Ward, the population was increasing, and those who were new in the community were appointed pieces of land to cultivate and to build their homestead, at the expense of areas of veld, bushland or woodland. The question is: how much communal land does a community need to provide for itself, and could other land uses than ‘communal land’ meet these needs to a more or a lesser extent than the ‘communal lands’ do? Our data are not sufficient to answer these pressing questions, but we can venture into an analysis based on what we do know. From our own data, it is clear that many of the key NTFRPs can be substituted. Some villages in Ushe had collectively planted woodlots of the fast-growing exotic Eucalyptus spp. to provide themselves with poles and firewood. One farmer even had his own Eucalyptus plantation, and he said he was never short of firewood. No farmer in our sample had planted indigenous fruit trees, but there were many of them standing in and around the fields or homesteads. This is because indigenous fruit trees are selectively retained when woodland was cleared for creating new fields or homesteads (Campbell, 1987). However, the planting and growing of miombo fruit trees may also be a future option to ensure a sufficient supply of indigenous fruits (Akinnifesi et al. , 2006; Akinnifesi et al. , 2008). Wild vegetables were mainly collected from gardens and fallow fields, and the same was true for insects. But a number of insect species (such as caterpillars) would dwindle if woodland was lost (Mbata et al. , 2002). Wild animals, such as hares and antelopes, would almost certainly become very rare (Jenkins et al. , 2002), so they would have to be substituted by domesticated animals. This requires investments in the form of cash, labour and feed. The woodlands also provide medicines, such as tree bark, that would be expensive to substitute. Here, again, important trees can be selectively retained while others are cut. Also, medicine is required only in small quantities, so smaller woodland areas, for example on mountains, would probably fulfill these needs. All in all, when looking at woodlands only from the ‘product’ perspective, they seem relatively easy to substitute. What is not easily substituted, however, is the role of the woodlands and rangelands in the overall system. In our sample, 20 out of 25 households used fertiliser, 16 out of 25 used manure and 15 out of 25 used leaf litter. Whereas the nutrients in fertiliser come

83 from outside the farming system, all nutrients from manure, leaf litter, compost and such come from within the systems. If these nutrients are no longer available because woodlands and rangelands are converted to fields, then farmers will have to substitute them by a form of fertiliser. Also, the free feed that is provided by woodlands and rangelands will have to be substituted by cultivated or purchased alternatives. This dependency on purchased inputs will make farmers more vulnerable to cash deficits and, thus, to poverty. And then, we have not even mentioned the importance of woodlands and rangelands as water catchment areas, for prevention of erosion, for cultural purposes, etc. (Kepe, 2007; Hough, 1986). The whole livelihood system of the farmers in our study was interwoven with the use of communal lands, even though substitutes for some NTFRPs were or could be made available. In our opinion, preservation of communal rangelands and woodlands is essential for the welfare of the smallholder farmers. Our results are insufficient to answer the question of ‘how much’ woodland must be preserved, but the quantitative data along with the land use maps that we made can contribute to the construction of an overall land use model for our study areas. Hartter and Boston (2007) constructed a mathematical model to study land use in relation to firewood harvesting by smallholder farmers in sub-Saharan Africa. This model, which is an adaptation of the FLORES model (Vanclay, 2003) allowed them to propose a minimum land requirement for a model community based on firewood use, alone. The expansion of this or similar models with other NTFRPs is required to calculate the minimum land requirements for our study area, especially in relation to climate change and variability. Even though we cannot calculate the minimum land requirements, in terms of areas, with the data that we have now, we can formulate qualitative requirements for land use. We propose four qualitative conditions which the communal natural areas must meet in order to provide the farmers in our study area with a sustainable natural resource basis for NTFRP harvesting. First, the woodlands and rangelands must be spatially distributed, to allow all farmers access to an area of woodland or rangeland within reasonable distance, preferably no further than the next village. We have observed that farmers generally choose to stay near to their homestead for the collection of NTFRPs, and additionally, most of the respondents in our sample indicated that they were short in labour. Therefore, it is a condition that the natural areas are equally distributed and within reach for all. Second, the areas must be in good condition; bush-encroached stretches of immature trees provide little resources for anyone. Especially in Ushe, farmers had bushlands nearby but there was hardly any firewood to be found because the trees were small and the area was intensively harvested. The yield of such areas is low, and they cannot provide the farmers with what they need. Third, the areas must be protected by a set of clear customary rules, in order to make their use sustainable. We found that in our study area, many farmers were either unaware of the rules, or they did not care to follow them. But without these rules, the communal lands are prone to degradation and their use cannot be sustainable. Therefore, re-establishment of the rules is required. Fourth, the range of areas within an administrative unit, in this case the ward or, for some NTFRPs, several wards, must be as diverse as possible. Each area (wetland, mountain, lowland woodland, veld, etc.) offers its own products, and farmers sample each selectively (when they are available) and are willing to walk long distances for products they highly value. The farmers from Ushe in our sample walked 20 kilometers there and 20 kilometer back to harvest U. kirkiana fruits. These four conditions are essential for sustainable natural resource use and a long- term contribution of NTFRPs to farmer’s welfare, at good and at bad times.

84 9.2 Conclusions

We have quantified the consumption of the most important NTFRPs in Wedza Communal Area, Zimbabwe, in order to see if, and how, the utilisation of these products can be a coping strategy for smallholder farmers at times of crop failure. Our data show that two wild fruits ( P. curatellifolia and U. kirkiana ) were consumed and/or collected at increased quantities in bad years. When looking at the overall energy availability, these wild fruits seemed to contribute a large share, especially at times of crop failure. Thus, our data agree with the hypothesis that NTFRPs are crucial to reduce food insecurity of smallholder farmers in Zimbabwe, both poor and wealthy. NTFRPs can play an especially large role in the light of the climatic uncertainties in the near future, because of the possible negative influences of climate change on food security. We strongly recommend that the resources in the communal areas are exploited with care, in accordance with the conditions that we formulated, so that they can be preserved to serve as a coping strategy at times of food insecurity.

85 86 References

Akinnifesi, F.K., Kwesiga, F., Mhango, J., Chilanga, T., Mkonda, A., Kadu, C.A.C., Kadzere, I., (...), Dhliwayo, P. Towards the development of miombo fruit trees as commercial tree crops in Southern Africa (2006) Forests Trees and Livelihoods , 16 (1), pp. 103-121.

Akinnifesi, F.K., Sileshi, G., Ajayi, O.C., Chirwa, P.W., Kwesiga, F.R., Harawa, R. Contributions of Agroforestry Research and Development to Livelihood of Smallholder Farmers in Southern Africa: 2. Fruit, Medicinal, Fuelwood and Fodder Tree Systems (2008) Agricultural Journal , 3 (1), pp. 76-88.

Alwang, J., Siegel, P.B. Labor shortages on small landholdings in Malawi: Implications for policy reforms (1999) World Development , 27 (8), pp. 1461-1475.

Belcher, B., Shreckenberg, K. Commercialisation of Non-timber Forest Products: A Reality Check (2007) Development policy review 25(3), pp 355-377.

Campbell, B.M., Luckert, M., Scoones, I. Local-level valuation of savanna resources: A case study from Zimbabwe (1997) Economic Botany , 51 (1), pp. 59-77.

Campbell, B.M. The Use of Wild Fruits in Zimbabwe (1987) Economic Botany , 41 (3), pp. 375-385.

Cavendish, W. Empirical regularities in the poverty-environment relationship of rural households: Evidence from Zimbabwe (2000) World Development , 28 (11), pp. 1979-2003.

Chavunduka, D.M. Insects as a source of protein to the African (1976) Rhod. Sci. News , 9, pp 217- 220.

Chenje, M., Sola, L., Paleczny, D. The State of Zimbabwe’s Environment 1998 (1998) Government of the Republic of Zimbabwe, Ministry of Mines, Environment and Tourism, Harare, Zimbabwe

Cocks, M.L., Bangay, L., Shackleton, C.M., Wiersum, K.F. Rich man poor man' - inter-household and community factors influencing the use of wild plant resources amongst rural households in South Africa (2008) International Journal of Sustainable Development and World Ecology 15 (3), pp. 198-210

Department of the Surveyor-General. Provisional Soil Map of Zimbabwe Rhodesia (1979) Surveyor- General, Salisbury, Zimbabwe Rhodesia , edition 2.

Dovie, D.B.K., Shackleton, C.M., Witkowski, E.T.F. Conceptualizing the human use of wild edible herbs for conservation in South African communal areas (2007) Journal of Environmental Management , 84 (2), pp. 146-156.

Dovie, D.B.K. Whose involvement? - Can hierarchical valuation scheme intercede for participatory methods for evaluating secondary forest resource use? (2003) Forest Policy and Economics , 5 (3), pp. 265-283.

FAO. Hunger Portal (2010). http://www.fao.org/hunger/faqs-on-hunger/en/#c41476

FAO. World Programme for the Census of Agriculture (2010). http://www.fao.org/docrep/009/a0135e/a0135e00.htm#TOC

FAO. Country Profile: Food Security Indicators, Zimbabwe (2009). http://www.fao.org/fileadmin/templates/ess/documents/food_security_statistics/country_profiles/eng/Zi mbabwe_E.pdf

FAO. The State of Food Insecurity in the World (2009).

FAO. Animal Feed Resources Information System: Piliostigma thonningii (?)

87 http://www.fao.org/AG/Aga/AGAP/FRG/afris/Data/271.HTM

FAO. Non-Wood Forest Products Programme (2008). http://www.fao.org/forestry/nwfp/en/

FAO. Irrigation in Africa in figures: AQUASTAT survey 2005 (2005) Food and Agriculture Organisation. Rome.

FAO. Livestock systems research manual - Volume 1 (1990) ILCA Working Paper - 1

FAO. Food Composition Table for Use in Africa (1968). http://www.fao.org/docrep/003/X6877E/X6877E00.htm#TOC

FAO. Animal Feed Resources Information System (no date). http://www.fao.org/ag/AGa/AGAP/FRG/AFRIS/Data/271.HTM

Ferro-Luzzi, A., Morris, S.S., Taffesse, S., Demissie, T., D'Amato, M. Seasonal undernutrition in rural Ethiopia: magnitude, correlates, and functional significance (2002) Food and Nutrition Bulletin , 23 (2), pp. 227-228.

Frost, P. The ecology of Miombo woodlands (1996) The Miombo in transition: Woodlands and welfare in Africa , pp. 11-55. In: Campbell, B. (Ed.) CIFOR, Bogor, Indonesia, Chapter 2.

Goebel, A., Campbell, B., Mukamuri, B., Veeman, M. People, values, and woodlands: A field report of emergent themes in interdisciplinary research in Zimbabwe (2000) Agriculture and Human Values , 17 (4), pp. 385-396.

Gomez, M.I. A resource inventory of indigenous and traditional foods in Zimbabwe (1988) Zambezia , XV (i), pp. 53-73.

Grundy, I., Turpie, J., Jagger, P., Witkowski, E., Guambe, I., Semwayo, D., Solomon, A. Special section: land use options in dry tropical woodland ecosystems in Zimbabwe. Implications of co- management for benefits from natural resources for rural households in north-western Zimbabwe (2000) Ecological Economics , 33, pp. 369–381.

Grundy, I.M., Campbell, B.M., Balebereho, S., Cunliffe, R., Tafangenyasha, C., Fergusson, R., Parry, D. Availability and use of trees in Mutanda Resettlement Area, Zimbabwe (1993) Forest Ecology and Management , 56 (1-4), pp. 243-266.

Guinand, Y., Lemessa, D. Wild food plants in Ethiopia. Reflections on the role of ‘wild-foods’ and ‘famine-foods’ at a time of drought (2001) UN Emergency Units for Ethiopia.

Hallund, J., Hatlöy, A., Benesi, I., Thilsted, S.H. Snacks are important for fat and vitamin intakes among rural African women: A cross-sectional study from Malawi (2008) European Journal of Clinical Nutrition 62 (7), pp. 866-871.

Hartter, J., Boston, K. An integrated approach to modeling resource utilization for rural communities in developing countries (2007) Journal of Environmental Management , 85 (1), pp. 78- 92.

High, C., Shackleton, C.M. The comparative value of wild and domestic plants in home gardens of a South African rural village (2000) Agroforestry Systems , 48 (2), pp. 141-156.

Hough, J. Management alternatives for increasing dry season base flow in the miombo woodlands of southern Africa (1986) Ambio , 15 (6), pp. 341-346.

Hulme, M., Doherty, R., Ngara, T., New, M., Lister, D. African climate change: 1900-2100 (2001) Clim Res , 17, pp 145-168.

88 IDRC. Resilience and the African Smallholder : Enhancing the Capacity of Communities to Adapt to Climate Change (2007) http://www.idrc.ca/en/ev-83071-201_104140-1-IDRC_ADM_INFO.html

Illgner, P., Nel, E. The geography of edible insects in Sub-Saharan Africa: A study of the mopane caterpillar (2000) Geographical Journal , 166 (4), pp. 336-351.

Jenkins, R.K.B., Corti, G.R., Fanning, E., Roettcher, K. Management implications of antelope habitat use in the Kilombero Valley, Tanzania (2002) ORYX , 36 (2), pp. 161-169.

Kepe, T. Beyond the numbers: Understanding the value of vegetation to rural livelihoods in Africa (2008) Geoforum , 39 (2), pp. 958-968.

Kinsey, B., Burger, K., Gunning, J.W. Coping with drought in Zimbabwe: Survey evidence on responses of rural households to risk (1998) World Development , 26 (1), pp. 89-110.

Le Houérou, H.N. Climate change, drought and desertification (1996) Journal of Arid Environments , 34, pp 133-185.

López, S., Davies, D.R., Giráldez, F.J., Dhanoa, M.S., Dijkstra, J., France, J. Assessment of nutritive value of cereal and legume straws based on chemical composition and in vitro digestibility (2005) J Sci Food Agric , 85, pp 1550-1557.

Luckert, M.K., Wilson, J., Adamowicz, V., Cunningham, A.B. Household resource allocations in response to risks and returns in a communal area of western Zimbabwe (2000) Ecol. Econ. 33, 383–394.

MacIntyre, U.E., Kruger, H.S., Venter, C.S., Vorster, H.H. Dietary intakes of an African population in different stages of transition in the North West Province, South Africa: The THUSA study (2002) Nutrition Research , 22 (3), pp. 239-256.

Maphosa, B. Lessons from the 1992 Drought in Zimbabwe: the Quest for Alternative Food Policies (1994) Nordic Journal of African Studies , 3 (1), pp 53-58.

Mazvimavi, D. Investigating possible changes of extreme annual rainfall in Zimbabwe (2008) Hydrology and Earth System Sciences Discussions 5, pp 1765-1785.

Mbata, K.J., Chidumayo, E.N., Lwatula, C.M. Traditional regulation of edible caterpillar exploitation in the Kopa area of Mpika district in northern Zambia (2002) Journal of Insect Conservation , 6 (2), pp. 115-130.

McGregor, J. Gathered produce in Zimbabwe's communal areas: changing resource availability and use (1995) Ecology of Food & Nutrition , 33 (3), pp. 163-193.

Mithöfer, D. Economics of Indigenous Fruit Tree Crops in Zimbabwe (2004) PhD Thesis, Department of Economics and Business Administration, University of Hannover, Hannover, Germany.

Mithöfer, D., Waibel, H. Income and labour productivity of collection and use of indigenous fruit tree products in Zimbabwe (2003) Agroforestry Systems , 59 (3), pp. 295-305.

Moyo, S. Land and natural resource redistribution in Zimbabwe: Access, equity and conflict (2005) African and Asian Studies , 4 (1-2), pp. 187-224.

Mugabe, F.T., Hodnett, M.G., Senzanje, A. Effect of temporal rainfall distribution and soil type soil moisture and runoff generation in semi-arid Zimbabwe (2007) Nordic Hydrology 38 (3), pp. 249- 263

Muller, J., Almedom, A.M. What is "famine food"? Distinguishing between traditional vegetables and special foods for times of hunger/scarcity (Boumba, Niger) (2008) Human Ecology , 36 (4), pp. 599-607.

89

Ncube, B., Twomlow, S.J., Dimes, J.P., Van Wijk, M.T., Giller, K.E. Resource flows, crops and soil fertility management in smallholder farming systems in semi-arid Zimbabwe (2009) Soil Use and Management , 25 (1), pp. 78-90

Ngwerume, F.C., Mvere, B. Report on the findings of a socio-economic survey on the marketing, consumption and production, of traditional vegetables in the urban and peri urban areas of Harare, Zimbabwe (2000) DR&SS-HRC/NRI Project A0892

O’Flaherty, M. The Tragedy of Property: Ecology and Land Tenure in Southeastern Zimbabwe (2003) Human Organization , 62 (2), pp. 178-190.

O’Flaherty, M. Communal Tenure in Zimbabwe: Divergent Models of Collective Land Holding in the Communal Areas (1998) Africa: Journal of the International African Institute 68 (4), pp 537-557.

Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D'Amico, J.A., (...), Kassem, K.R. Terrestrial ecoregions of the world: A new map of life on Earth (2001) BioScience , 51 (11), pp. 933-938.

Paumgarten, F., Shackleton, C.M. Wealth differentiation in household use and trade in non-timber forest products in South Africa (2009) Ecological Economics 68 (12), pp. 2950-2959

Paumgarten, F. The role of non-timber forest products as safety-nets: A review of evidence with a focus on South Africa (2005) GeoJournal, 64 (3), pp. 189-197.

Pearson, R.A. Pastures: the overstocking issue. Feeding strategies for cattle and buffalo used for work (1996) World Animal Review , 87. http://www.fao.org/docrep/W2650T/w2650t05.htm

Peters, C.M., Gentry, A.H., Mendelsohn, R.O. Valuation of an Amazonian rainforest (1989) Nature , 339 (6227), pp. 655-656.

Phillips, J.G., Cane, M.A., Rosenzweig, C. ENSO, seasonal rainfall patterns and simulated maize yield variability in Zimbabwe (1998) Agricultural and Forest Meteorology , 90 (1-2), pp. 39-50.

Saka, J.D.K., Msonthi, J.D. Nutritional values of edible fruits of indigenous wild trees in Malawi (1994) Forest Ecology and Management , 64, pp. 245-248.

Scoones, I., Chibudu, C., Chikura, S., Jeranyama, P., Machaka, D., Machanya, W., Mavedzenge, B., Mombeshora, B., Mudhara, M., Murimbarimba, F. and Zirereza, B. Hazards and Opportunities, Farming livelihoods in Dryland Africa: Lessons from Zimbabwe (1996) Zed Books, London.

Scoones, I. Exploiting heterogeneity: Habitat use by cattle in dryland Zimbabwe (1995) Journal of Arid Environments , 29 (2), pp. 221-237.

Scoones, I.C., Melnyk, M., Pretty, J. The Hidden Harvest: Wild Foods and Agricultural Systems. A Literature Review and Annotated Bibliography (1992) 1.11.1. ODI, London.

Shackleton, C.M., Shackleton, S.E. Household wealth status and natural resource use in the Kat River valley, South Africa (2006) Ecological economics 57, pp 306-317.

Shackleton, C.M., Shackleton, S.E. The importance of non-timber forest products in rural livelihood security and as safety nets: A review of evidence from South Africa (2004) South African Journal of Science , 100 (11-12), pp. 658-664.

Shackleton, S.E., Shackleton, C.M., Netshiluvhi, T.R., Geach, B.S., Ballance, A., Fairbanks, D.H.K. Use patterns and value of savanna resources in three rural villages in South Africa (2002) Economic Botany , 56 (2), pp. 130-146.

Surveyor-General Zimbabwe. 1:50.000 map of ?, Zimbabwe (1974) Zimbabwe.

90 Turner, M.D., Hiernaux, M., Schlecht, E. The Distribution of Grazing Pressure in Relation to Vegetation Resources in Semi-Arid West Africa: The Role of Herding (2005) Ecosystems , 8 (6), pp. 668-681.

Twine, W., Moshe, D., Netshiluvhi, T., Siphugu,V. Consumption and direct-use values of savanna bio-resources used by rural households in Mametja, a semi-arid area of Limpopo province, South Africa (2003) South African Journal of Science 99, pp. 467–474.

Tyynelä, T., Niskanen, N. Use and Sustainability of Miombo Woodlands under Community Management in Zimbabwe (2009) Nordic Journal of African Studies , 9(1), pp. 118-142.

United Nations. United Nations Millennium Declaration (2000) Declaration of the fifty-fifth session of the General Assembly , www.un.org/millennium

Usman, M.T., Reason, C.J.C. Dry spell frequencies and their variability over southern Africa (2004) Climate Research , 26 (3), pp. 199-211.

Vanclay, J.K. Why model landscapes at the level of households and fields? (2003) Small-Scale Forestry , 2 (2), pp. 121-134.

Vermeulen, S.J. Cutting of trees by local residents in a communal area and an adjacent state forest in Zimbabwe (1996) Forest Ecology and Management , 81 (1-3), pp. 101-111.

Vincent, V., Thomas, R.G. An agricultural survey of Southern Rhodesia, Part I: Agro-ecological survey (1960) Government Printer, Salisbury, Rhodesia.

Vorster, H.H., Kruger, A., Margetts, B.M., Venter, C.S., Kruger, H.S., Veldman, F.J., MacIntyre, U.E. The nutritional status of asymptomatic HIV-infected Africans: Directions for dietary intervention? (2004) Public Health Nutrition , 7 (8), pp. 1055-1064.

Wolmer, W., Scoones, I. The science of 'civilized' agriculture: The mixed farming discourse in Zimbabwe (2000) African Affairs , 99 (397), pp. 575-600.

Women and Men of Zimbabwe Arise! Looking back to look forward - education in Zimbabwe: a WOZA perspective (2010) WOZA Zimbabwe, http://wozazimbabwe.org/

Zingore, S., González-Estrada, E., Delve, R.J., Herrero, M., Dimes, J.P., Giller, K.E. An integrated evaluation of strategies for enhancing productivity and profitability of resource-constrained smallholder farms in Zimbabwe (2009) Agricultural Systems 101 (1-2), pp. 57-68

Zinyama, L.M., Matiza, T., Campbell, D.J. The use of wild foods during periods of food shortage in rural Zimbabwe (1990) Ecology of Food and Nutrition 24, pp 251-265.

91 92 Appendix A: Conversion Factors

A1: CONVERSION FACTORS FOR FARM PRODUCTS Farm product Quantity Conversion Source Maize 1 bag 50 kg DW According to respondents Potatoes 1 litre 769 g DW http://www.simetric.co.uk/si_materials.htm Soybean 1 litre 721 g DW http://www.simetric.co.uk/si_materials.htm Groundnut (unshelled) 1 litre 272 g DW http://www.simetric.co.uk/si_materials.htm Beans 1 litre 801 g DW http://www.simetric.co.uk/si_materials.htm Wheat flour 1 litre 593 g http://www.simetric.co.uk/si_materials.htm Millet grain 1 litre 780 g DW http://www.simetric.co.uk/si_materials.htm Rice 1 litre 753 g DW http://www.simetric.co.uk/si_materials.htm Wheat grain 1 litre 790 g DW http://www.simetric.co.uk/si_materials.htm Bambara groundnut 1 litre 750 g DW Own estimate Pumpkin 1 piece 1 kg FW Own estimate 1 cart 200 kg FW Own estimate Butternut 1 piece 0.5 kg FW Own estimate 1 cart 200 kg FW Own estimate 1 bucket 10 kg FW Own estimate Kale/rape vegetables 1 bundle 0.5 kg FW Own estimate Sugarcane 1 stalk 1 kg FW Own estimate Bananas 1 bunch 5 kg FW Own estimate http://www.whfoods.com/genpage.php?tname=foodspi Sunflower seed 1 bag 30 kg FW ce&dbid=57 Tomatoes 5 litre 2.72 kg FW http://wiki.answers.com Cucumber 1 bucket 12 kg FW Own estimate Onion 1 bucket 10.88 kg FW Own estimate Carrots 1 bucket 10.88 kg FW Own estimate Oranges 1 bucket 10.88 kg FW Own estimate Mango 1 bucket 12 kg FW Own estimate Guava 1 bu 10.88 kg FW Own estimate Pawpaw 1 fruit 0.5 kg FW Own estimate Grape 1 bunch 1 kg FW Own estimate Apple 1 fruit 150 g FW Own estimate Spring onion 1 piece 20 g FW http://www.answers.com/topic/spring-onion-1 Cabbage 1 head 0.5 kg FW Own estimate Chicken 1 chicken 1.5 kg meat http://wiki.answers.com Cattle 1 cattle 144 kg meat ?? Turkey 1 turkey 4 kg meat Own estimate Goat 1 goat 30 kg meat Own estimate

93

A2: CONVERSION FACTORS FOR NTFRPs NTFRP Quantity Conversion Source Firewood 1 bundle 25 kg DW 1 bundle 27 kg DW Biran, Abbot & Mace, 2004 1 bundle 25 or 26 kg DW http://www.fao.org/docrep/Q1085E/q1085e0c.htm 1 bundle 17.5 kg DW Tabuti, Dhillion & Lye, 2003 1 bundle 29.9 kg DW Ham, 2000 1 cartload 200 kg DW Hyman 1982 1 cartload 340 kg DW Benjaminsen 1997 1 cubic m 600 kg DW http://www.fao.org/DOCREP/006/AD600E/ad600e00.pdf Parinari curatellifolia 1 litre 0.6 kg FW Own measurements 1 fruit 15 g FW Own measurements Uapaca kirkiana 1 liter 0.6 kg FW Own measurements 1 fruit 15 g FW Own measurements http://www.worldagroforestry.org/downloads/publications/ 1 fruit 5 - 50 g FW PDFs/pp05177.doc Amanita zambiana 1 litre 250 g FW http://www.veg-world.com/articles/cups.htm Thatch grass 1 bundle 0.85 kg FW http://www.sa-thatchers.co.za/ Flying termites 1 litre 208 g FW Own estimate Poles 1 pole 10 kg DW http://www.africanwoods.co.za/products.html 1 cartload 20 poles DW Own estimate http://www.biodiversityexplorer.org/mammals/lagomorph Cape hare 1 hare 1.5-4.5 kg FW a/lepus_saxatilis.htm 1 hare 3 kg FW Own estimate Strychnos spinosa 1 fruit 0.3 kg FW Own measurements 1 litre 0.8 kg FW Own estimate Grasshoppers 1 adult 2 g FW http://en.wikipedia.org/wiki/Migratory_locust 1 adult 2.5 g FW Own estimate 1 litre 0.6 kg FW Own estimate 1 litre 705 g DW http://www.simetric.co.uk/si_materials.htm Vitex payos 1 litre 0.6 kg FW Own estimate Azanza garckeana 1 litre 0.5 kg FW Own estimate Own estimate, based on Leaf litter 1 cubic m 210 kg DW http://www.simetric.co.uk/si_materials.htm Dicoma anomala 1 litre 0.8 kg FW Own estimate Birds 1 bird 25 g FW Own estimate Cochorus oditorius 1 cup 75 g FW Own estimate, based on http://wiki.answers.com Coleus esculentus 1 litre 325 g FW Own estimate

94

A3: CONVERSION FACTORS LOCAL UNITS TO UNIVERSAL UNITS Local unit Quantity Conversion Source Scotch cart 1 cart 800 l Own measurements Wheelbarrow 1 barrow 27.5 l Own measurements Handful 1 handful 150 ml Own estimate Plate 1 plate 0.7 l Own estimate Deep plate 1 plate 1.5 l Own estimate Cup 1 cup 250 ml Own estimate Big cup 1 cup 500 ml Own estimate Small dish 1 dish 3 l According to respondents Dish 1 dish 5 l Own estimate Bucket 1 bucket 20 l According to respondents Bag 1 bag 50 kg of maize According to respondents 1 bag 69 l http://www.simetric.co.uk/si_materials.htm 90 kg bag 1 bag 90 kg of maize According to respondents

A4: OTHER CONVERSIONS Other Quantity Conversion Source "Almost every day" 6 days/week

95 96 Appendix B: Household questionnaire

Questionnaire version 6 (revised on 14/10/09) Dates:

Part 1: General Household Characterisation

1a Names (full) of respondents

1b Position of respondents, in relation to Household Head

1c Name of village

1d Position of the homestead (GPS/map)

1e Further household characterisation Name Relation with HH Age Education

1f How big is your farm?

1g How big is your garden?

1h MAPPING of the farm with the homestead, the different fields and the garden.

1i What livestock do you have? Type Number

1j Who are working on your farm?

1k Is there enough labour to do all the farming activities that you want?

1l Do you hire labour? If yes: how much, what for, and how much do you pay for it?

97 Part 2: Farm products

2a What crops did you grow this season, on what fields (MAP), when did you harvest them and what was their yield?

2b What crops did you grow this season in the garden (including fruits), when did you harvest them and what was their (estimated) yield? Crop Field/garden Yield Time of harvesting

2c What inputs did you apply to which crops, and how much of them? Input Quantity Crop

2d What animal products did you produce last year (since October 2009), when did you 'harvest' them and what was their yield? Product Yield Time of harvesting

98 Part 3: consumption in normal years

3a How much of your produce did you keep for own consumption, and when did you consume it? Product Consumed Stored Period of consumption

3b How much of your produce (including fruits and animal products) did you give away/sell/barter, and what were the returns? Product Given away Sold/bartered Returns

3c Did you buy any food since October last year? If yes: what, when and how much? Product Period of purchase Quantity

3d Do you have any other sources of income than your farming activities? If yes: which, and how much do they contribute?

3e What did you eat last week? Dish Main ingredients Quantity per week Source

99 Part 4: livestock management

4a What do you feed your livestock in the different periods of the year, and how much? Livestock type Period of the year Type of feed Quantity

4b Where does your livestock graze in the different months of the year (indicate on map)?

4c How do you select the grazing areas for your livestock?

100 Part 5: Flood or drought years

5a Can you describe the weather situation in the last cropping season?

5b What year do you remember best in which the weather was very bad for cropping?

5c What was the weather like in this year?

5d What crops (including garden crops and fruits) did you plant in this year, and what was their yield? Crop Yield

5e What animal products did you produce this year and what was their yield? Product Yield Time of harvesting

5f How much of your produce did you keep for own consumption, and when did you consume it? Product Consumed Stored Period of consumption

5g How much of your produce (including fruits and animal products) did you give away/sell/barter, and what were the returns? Product Given away Sold/bartered Returns

5h What other sources of food or income did you have during this years?

101

5i Did you buy or were you given any food in this year? If yes: what, when and how much? Product Period of purchase Quantity

5j What did you eat in a week in October, on average? Dish Main ingredients Quantity per week Source

5k What did you feed the livestock in the different periods of the year? Livestock type Period of the year Type of feed Quantity

5i Did you change your herding strategy in this year (grazing routes, nr of days of herding)? If yes, how (indicate on the map)?

102 Part 6: Extraction tables 6a Do you gather any of the following products from the common lands? Product Names

Fruits Herbs/vegetables Roots/tubers Mushrooms Insects Animals Firewood Grazing Animal feed Construction poles Leaf litter Termitaria Other

Top 5 of most important products (food AND non-food) Normal years: Bad years:

103 FOOD PRODUCTS

5b In what months can you collect this product?

6a In these months: how often do you consume the product?

6b How much of the product do you consume each time?

6c Why do you consume this product?

6d Who consumes/uses the product?

6e How often is the product collected and how much is collected each time?

6f Who collects the product and how much time does the collection take?

6g Is the product stored (how long) or consumed immediately?

6h Is the product sold or bartered? If yes, how much is sold/bartered and at what price?

6i What is the precise use of the product, how is it used/prepared, what are possible substitutes?

6j Do you think there is enough of the product available? Is the availability different from 10 years ago? If yes: how?

7. NON-FOOD PRODUCTS

7a When is this product collected?

7b In this period: how often is the product collected, and how much is collected each time?

7c Who goes to collect the product and how much time does the collection take?

7d Is the product sold or bartered?

7e How is the product used?

104 Part 8: Mapping Composition of a map, showing: - Homestead - Main roads, rivers, mountains, etc. - Main area of common land that are utilised by the household (characterise vegetation type) - Areas of restricted access - Nr of resources available (eg nr of tree) - Grazing routes in normal vs dry years

Part 9: Access - Which are areas of restricted access? - Where is the cattle allowed to graze? - Where are you allowed to collect firewood? - Where are you allowed to collect wild fruits/animals?

105 106 Appendix C: Access (interview with the headman of Ushe Ward)

In order to get a better picture of the rules and regulations regarding the collection and use of NTFRPs, I interviewed Mr. Ushe, the headman of Ushe Ward. He had been in office for two years, and he had inherited the title from his father. We discussed topics ranging from the use of indigenous food plants to the political structure in rural Wedza. Below is a summary of our discussion. The title of headman is inherited from father to son. The headman is the highest leader of the ward. In addition, each ward has a counsellor, whose function is more administrative. The counsellor is elected by the people in the ward, and he represents the ward in the district council. In general, the custom is that indigenous foods, especially the fruits and plants, are available to anyone who needs them. During the 2007/2008 drought, the collection of mobola plum and wild loquat in other wards was negotiated with the villagers there. The only problem of collecting in other wards was when people were collecting from trees in yards or homesteads. In some cases, when people have trees in their yard, they can be only picked after asking permission, but according to the customary law this permission cannot be denied if the fruit tree is indigenous. So it is mere politeness to ask permission, but indigenous fruit trees cannot be 'owned' by someone. Also it is not allowed, according to customary law, to sell any indigenous fruits. However, it does happen that townspeople come into the rural areas to collect fruits for selling them in town. Kraalheads control the obedience to customary law within their kraal, but still illegal harvesting happens, especially at night. If the police catches a truck loaded with indigenous fruits on a road block, the driver is prosecuted, which means that customary law and national law are similar regarding the harvesting of indigenous fruits. To summarise, everyone is allowed to use the indigenous food plant resources whenever hungry, but it is not allowed to sell them. Kraalheads and even the national police enforce this customary law. Dried firewood, either wood that is already on the ground or dead branches that are still on the tree, can be collected freely by kraal members within the kraal boundaries. To collect firewood in another kraal, one must ask permission from the kraalhead. This is common practice, as some kraals have much more woodland than others. For cutting trees, permission is required from the kraalhead, who reports the cutting to the headman and the counsellor. However, it is the kraalhead who decides whether or not a tree can be cut. Trees on the mountains may never be cut, for reasons of erosion, security (e.g. as a hiding place in events of war) and tradition. The mountains are for example used in the rituals of spirit mediums and the trees on the mountains are the domain of the ancestor spirits and must be preserved. People from other wards are not allowed to come into the ward to cut or collect firewood. All animals may be hunted, according to customary law. There are no rules about numbers of animals that can be hunted, but currently the animal resources are not being over-exploited, according to the headman. Hunting in other wards is not allowed according to customary law. According to national law, no hunting at all is allowed, and if officers catch community members carrying a catapult (the most common hunting weapon) they will take it from them. However, there are usually no officers around. In the past, cattle grazing was restricted to grazing areas. Nowadays, cattle can be grazed in most of Ushe Ward, ignoring kraal boundaries. The kraalhead can appoint some areas that are prohibited for cattle grazing, and fields under cropping are always

107 prohibited. It is not allowed to graze cattle in other wards, but in very special cases, permission can be asked from the ward headman. Kraal boundaries coincide with natural boundaries such as streams, low-lying wetland regions and mountains. The boundary areas are shared and used by all kraals that lie adjacent to it. For example, if a mountain forms the boundary of the kraal, then people from both kraals can use this mountain freely. Sometimes farmers ignite veld-fires as a form of management, but according to customary (and national) law this is not allowed. The headman holds court every Friday, and sometimes cases of illegal use of natural resources come up. In that case, the headman can demand a fine, the height of which is based (in Ushe Ward) on the carrying capacity of the person being fined. If no cash is available, the fine will be paid in the form of a chicken or a goat. A goat is the highest possible fine that is given. The fines become property of the headman, but they do not make him rich because in general, people obey very well to the customary laws of the ward. The population of Ushe Ward is now growing. Immigration into the ward is allowed, provided that there is a place in one of the kraals. This is decided upon by the kraal head.

108