Human Ecology https://doi.org/10.1007/s10745-018-9981-2
The Influence of Soil Quality and Market Orientation on Manioc (Manihot esculenta) Varietal Choice by Smallholder Farmers along the Lower Tapajós River, Pará, Brazil Raquel Sousa Chaves 1 & André Braga Junqueira 2,3,4 & Charles R. Clement 5
# Springer Science+Business Media, LLC, part of Springer Nature 2018
Introduction Manioc (Manihot esculenta Crantz) is the main crop cultivated by traditional family farmers in Amazonia, generally managed in shifting cultivation systems comprising slash-andburn clearing of plots, cropping for a year or so, followed by a fallow period. Besides restoring soil nutrients and organic matter, fallows are an important source of food and other resources (Junqueira et al. 2010). Manioc has great economic and cultural importance, especially for traditional and indigenous communities who cultivate large numbers of local varieties (Emperaire and Eloy 2008; Heckler and Zent 2008; Fraser et al. 2012; Lima et al. 2012; Robert et al. 2012; Lima et al. 2013; Peña-Venegas et al. 2014; Junqueira et al. 2016a). The management systems and the diversity of manioc varieties cultivated are influenced by ecological, socioeconomic, and cultural factors, such as the type of soil (Fraser and Clement 2008; Fraser et al. 2011a), soil fertility (Junqueira et al. 2016a), culinary preferences, ethnicity (Peña-Venegas et al. 2014), and market orientation (Peroni and Hanazaki 2002; Lima et al. 2012). Orientation towards * Raquel Sousa Chaves
[email protected] 1
Programa de Pós-graduação em Botânica, Instituto Nacional de Pesquisas da Amazônia, Av. André Araujo 2936 - Petrópolis, Manaus, Amazonas 69067–375, Brazil
2
International Institute for Sustainability, Rio de Janeiro, Brazil
3
Centre for Conservation and Sustainability Science (CSRio), Department of Geography and the Environment, Pontificial Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
4
Department of Soil Quality, Wageningen University and Research, Wageningen, The Netherlands
5
Instituto Nacional de Pesquisas da Amazônia, Manaus, Amazonas 69067–375, Brazil
the market may lead to a gradual change from varieties used for home consumption to those that meet the demands of the market, leading to a decrease in agrobiodiversity (Peroni and Hanazaki 2002; Kumar and Nair 2004; Peyre et al. 2006) and to an increase in the area cultivated with Bcommercial varieties^ (Major et al. 2005). Soils in Amazonia are generally acidic and nutrient poor (EMBRAPA 2013), but there are also patches of Amazonian Dark Earths (ADE), locally called Terra Preta, which are fertile anthropogenic soils created through the accumulation of organic materials in ancient indigenous settlements (Glaser and Birk 2012). These soils are rich in phosphorus, calcium, magnesium, and charcoal, and have higher pH than background soils (Falcão et al. 2009; Glaser and Birk 2012). In shifting cultivation systems along the Madeira River (Central Amazonia), ADE are associated with more intensive manioc cultivation, shorter cropping cycles, shorter fallow periods, and higher frequency of cultivation (Fraser et al. 2011a; Junqueira et al. 2016b). Unlike other rivers in Amazonia, the Tapajós River has clear water and no fertile floodplains enriched by sediments from the Andes. The geology of the lower Tapajós (EMBRAPA 2013), which influences the characteristics of the anthropogenic and non-anthropogenic soils formed in the region, is different from that along whitewateror blackwater-rivers. Thus, it is important to establish whether the intensive manioc cultivation systems on ADE reported on the Madeira River are idiosyncratic to that region or are part of a more widespread trend in traditional Amazonia in regions with contrasting environmental backgrounds, such as the lower Tapajós. In Amazonia, particulaly in peri-urban areas, there is growing intensification of traditional agricultural systems owing to increasing demand and, in some cases, to shortages of accessible land for cultivation (Jakovac et al. 2017). Some traditional populations that reside inside the so-called ‘sustainable use conservation units’ also face stricter regulations on the
Hum Ecol
amount of land that can be cultivated per capita and on the use of fire in their cultivation systems (SNUC 2000). The lower Tapajós, like other areas in Amazonia, is gradually urbanizing (Ioris 2015) and market orientation is increasing (Vaz-Filho 1997). It is important, therefore, to investigate in detail the effect of these processes on cultivation systems, in order to plan and support, together with local farmers, strategies that foster the conservation of agrobiodiversity and increase food security of smallholder farmers. This study aimed to investigate the influences of soil types and fertility (associated with ADE) and orientation to the market on the choice of manioc varieties in the Tapajós-Arapiuns Extractive Reserve, lower Tapajós River.
Methods Study Area This study was conducted in three riverside communities located in the Tapajós-Arapiuns Extractive Reserve, on the left bank of the lower Tapajós River, Pará, Brazil: Enseada do Amorim (68 households), Parauá (227) and Surucuá (95) (ICMBIO 2014) (Fig. 1). The communities were selected because the first author is a resident of one of the communities (Surucuá), which facilitated access to the area and trust of the residents. The Tapajós-Arapiuns
Resex was created on 6 November 1998 in an area of 647,610 ha., with an estimated population of 23,000 inhabitants (about 4000 households) in 72 communities located on the banks of the Tapajós and Arapiuns Rivers, and to a lesser extent, along smaller rivers inside the reserve (ICMBIO 2014). The predominant soils in the region are distrophic red / yellow Oxisols, but red / yellow Ultisols and sandy Neosols also occur (IBGE 1976). Gomes (2008) described 10 archaeological sites on uplands in and around our study area that were initially occupied 3800–3600 BP by semi-sedentary groups and by sedentary groups between 1300 and 910 BP. These archaeological sites are associated with patches of ADE. Current communities are usually located on former indigenous villages or missionary settlements that attracted indigenous peoples. Most current communities formed after the Cabanagem (1835–1840), the war between the newly formed Brazilian state and indigenous peoples that had a profound effect on modern settlement patterns (Vaz-Filho 1997; Oliveira 2006). Houses are located on riverbanks, while agricultural fields are more distant from the rivers. Farming is the main source of income, with manioc the most important crop and its flour the main commercialized product. Along with manioc, most farmers also plant some maize (Zea mays L.), beans (Phaseolus vulgaris L.), squash (Cucurbita maxima Duch.), yams (Dioscorea spp.), sweet potato (Ipomoea
Fig. 1 Study are and research sites. White points represent the communities where the research was conducted in the Resex Tapajós-Arapiuns and the seat of the municipality of Santarém, lower Tapajós River, Pará, Brazil. Images obtained from Google Earth
Hum Ecol
batatas (L.) Lam.), chili peppers (Capsicum frutescens L., C. chinense Jacq.), gherkin (Cucumis anguria L.), watermelon (Citrullus lanatus (Thunb.) Matsum & Nakai), banana (Musa spp.), sugarcane (Saccharum spp.), and mangarataia (Zingiber officinale Roscoe). Residents also fish, hunt, and collect fruits for household consumption.
Data Collection and Analysis Thirty households participated in the study: eight from Enseada do Amorim, 14 from Parauá, and eight from Surucuá, selected through the Bsnowball technique^ of identification by community members (Albuquerque et al. 2014), and aimed at maximizing the variation between the households regarding commercial production of manioc flour and the types of soils in their fields. We sampled 61 fields in the three communities, a minimum of two swiddens per family; the total area cultivated by each family is on average 1.8 ± 1.0 ha. Data were obtained through semi-structured interviews, guided tours, and participant observation (Albuquerque et al. 2014). The interviews were conducted with several family members at the same time, which allows access to Bcollective memory^ (Halbwachs 1990). We solicited information on the size of each swidden, the duration of each cropping cycle (i.e., time to maturity), the number of successive cropping cycles in the same field, fallow duration, perceptions about the suitability of different soils for manioc production, the number of varieties, the area occupied by each variety in each field (as a proportion of the size of the field, according to the perception of the farmer), the performance and productivity of each variety, use of rare varieties, types of products produced and their destinations (market or home consumption), the amount of flour produced monthly per family (50 kg bags), and the amount of flour each household intended to market monthly. We use the term variety to refer not to a variety in the formal agronomic sense (i.e., subjected to formal genetic improvement), but to local definitions (Villa et al. 2005). Many of the interviewees categorized manioc varieties as ‘weak’ and/or ‘watery,’ and ‘strong’ and/or ‘dry,’ similar to categorizations reported in other regions in Amazonia (Fraser and Clement 2008; Lima et al. 2012). For each variety mentioned, the interviewee was asked to classify it on an ordinal scale starting from the weakest (score = 1) to the strongest variety. The mean value cited for a particular variety was considered its ‘strength index’ (Fraser and Clement 2008), which was used to investigate the association between the ‘strength’ of the variety and different types of soil already reported for the Madeira River (Fraser and Clement 2008; Fraser et al. 2011a). Fields were visited with informants in guided-tours during which manioc varieties, soil characteristics (color, texture, and presence or absence of ceramic fragments), and other species managed in the fields were identified, and the size of the fields
was measured. In each field, a composite soil sample (five subsamples; fields covered in zigzag) was collected from the 0–20 cm layer. Soil samples were dried, sieved, and submitted to physical (percentages of sand, clay and silt) and chemical analyses (pH (in water), pH (as potassium chloride), nitrogen (N), carbon (C), organic matter content (OM), available phosphorus (P), potassium (K), exchangeable calcium (Ca), magnesium (Mg), aluminum (Al), exchangeable acidity (H + Al), total iron (Fe), zinc (Zn) and manganese (Mn)). Analyses were performed in the Thematic Soil and Plant Laboratory at INPA, following EMBRAPA (1997). To visualize the variation in chemical and physical properties of the soils, a Principal Component Analysis (PCA) was performed after standardization of variables using the Vegan library (Oksanen et al. 2015) in R (R Core Team 2015). The frequency of varieties was used to calculate the number of shared and unique varieties among communities and soil types. To obtain the proportion of the area of each soil category that is cultivated with a specific variety, the sum of the area of a given variety in a given type of soil was divided by the total area of this soil. We also asked each interviewee to classify each variety according to the area they usually occupy in the fields (large or small areas) and to the number of households that usually cultivate them (many or few). This allowed comparisons of the areas occupied by the most common varieties in the fields, considering the different types of soil. The most frequent and abundant bitter varieties used for the production of yellow manioc flour primarily for the market were categorized as Bcommercial varieties.^ To compare the different soil types in terms of fallow duration and the number of successive cropping cycles, an analysis of variance (ANOVA) followed by a Tukey post-hoc test was used. The market orientation of each family was calculated as the proportion of manioc flour sold by the family in relation to the total amount produced. To test the effects of market orientation on the characteristics of the cultivation system, orientation to the market was used as a predictor in linear regressions, with the total size of the fields, the number of varieties grown and the proportion of the area cultivated with Bcommercial varieties^ as dependent variables. Variables with skewed distributions were transformed prior to the analyses (arcsin transformation for market orientation and the proportion of the area with Bcommercial varieties,^ and log transformation for the size of fields). All analyses were performed using R software (R Core Team 2015).
Results Diversity and Use of Manioc Varieties Forty-one varieties of manioc were found in the three communities. The farmers classified varieties into three groups
Hum Ecol
(Table 1): ‘mandioca’ – varieties with high concentrations of cyanogenic glucosides, described as bitter manioc in the
Table 1
literature; ‘macaxeira’ – varieties with low concentrations of cyanogenic glucosides, also known as sweet manioc; and
Manioc varieties cultivated by 30 families in three riverside communities in the Resex Tapajós-Arapiuns, lower Tapajós River, Pará, Brazil
Large areas/Many families
Large areas/Few families
Small areas/Many families Small areas/Few families
Variety name
Local classification
Root color (pulp)
Maturation (months)*
Strength index**
Total cultivated area (ha)
Relative frequency (families)
Acarizinha
Mandioca
Yellow
8
0.3
4.5
43.3
Achadinha Curuá-una
Mandioca Mandioca
Cream Yellowish
12 10
0.7 0.6
9.0 11.2
66.6 76.6
Mulata Pretinha-amarela
Mandioca Mandioca
Yellow Yellow
6 8
0.2 0.4
7.9 3.5
50 36.6
Boi
Mandioca
Yellow
12
0.5
1.6
10
Braço-trançado Enche-paneiro
Mandioca Mandioca
Yellow Yellowish
8 12
0.3 0.3
0.5 0.5
6.6 6.6
Inajazinha Maracanã
Mandioca Mandioca
Yellow Yellowish
8 12
0.3 0.5
0.6 0.8
10 6.6
Marrequinha Pinajé
Mandioca Mandioca
Yellow Cream
8 8
0.2 0.3
1.1 0.9
16.6 13.3
Pretinha-branca Seis-meses Viada Xingú Macaxeira-branca
Mandioca Mandioca Mandioca Mandioca Macaxeira
Cream Yellow Yellowish Yellow White
12 6 12 8 6
0.7 0.2 0.3 0.3 –
0.7 0.6 1.4 0.6 0.02
10 6.6 13.3 10 36.6
Macaxeira-boliviana Macaxeira Macaxeira-manteiga Macaxeira
White Yellow
6 6
– –
1.1 1.0
26.6 46.6
Apacê Arranha-céu Bonita Branquinha Brebeí
Mandioca Mandioca Mandioca Mandioca Mandioca
Yellow Yellowish White White Yellowish
12 12 10 12 12
0.4 0.7 0.5 0.8 0.3
0.1 0.3 0.3 0.1 0.4
3.3 6.6 10 3.3 3.3
Caetana Castanha Cobra Flecha
Mandioca Mandioca Mandioca Mandioca
Cream White Yellowish Yellow
12 12 12 8
0.4 0.5 0.8 0.3
0.3 0.4 0.1 0.3
3.3 6.6 3.3 13.3
Jamundá Ladrona Macaco Manicuera-branca Manicuera-preta Meciana Mete-medo Milagrosa Paraiso Peito-de-moça Preguiça Pretona Tartaruguinha
Mandioca Mandioca Mandioca Manicuera Manicuera Mandioca Mandioca Mandioca Mandioca Mandioca Mandioca Mandioca Mandioca
White Yellowish Yellow Yellowish Yellowish Yellow Yellow Yellow Yellowish Yellowish Yellow Cream Yellow
12 12 8 6 6 10 10 8 12 12 6 12 8
0.4 0.6 0.3 – – 0.4 0.3 0.4 0.2 0.5 0.2 1.0 0.2
0.1 0.3 0.2 0.02 0.03 0.1 0.1 0.1 0.5 0.1 0.1 0.5 0.4
3.3 3.3 6.6 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3
*Minimum time to maturation for each variety (consensus among farmers) **Mean strength attributed by farmers to each variety (from Fraser and Clement 2008)
Hum Ecol
‘manicuera’ – very watery varieties that produce a juice with high sugar content during processing (‘tucupí’). Bitter manioc varieties are used mainly in the production of manioc flour, known locally as ‘farinha d’água’ (‘water flour,’ referring to the fact that the fermentation process is done in water) or yellow flour, but also in the preparation of other products, such as ‘beijús’ (manioc flat cake), ‘carimã’ (mass used in the preparation of biscuits), ‘tarubá’ (a drink), among others. Among the 36 bitter varieties, 27 have yellowish root flesh color, five are cream colored and four are white (Table 1). The sweet manioc varieties are consumed after cooking (without fermentation), used in the preparation of cakes, beijús, and porridges. The manicueras are used in the preparation of a local beverage of the same name: the tucupí is extracted from the roots, and mashed pieces of sweet manioc and purple yam (Dioscorea trifida L.f.) are added. Seventeen varieties (41% of total) are cultivated by at least two communities, of which six are grown in all three communities, and 24 (59%) are grown exclusively in a single community (Fig. 2). Among the 41 varieties, eight are more frequently cultivated, managed by 8–23 households among the 30 households interviewed; five of these varieties are bitter, occupy proportionally larger areas in the fields, and are associated with production of flour for market (hereafter commercial varieties); three are sweet manioc varieties that occupy small areas in the fields and are also occasionally grown in homegardens (Table 1). The other 33 varieties are cultivated by at most five households - 11 occupy large areas in the fields and the remaining 22 small areas. It is noteworthy that among the latter are varieties with specific uses, such as two manicuera varieties, three varieties preferred for the preparation of tarubá (Cobra, Macaco, and Preguiça), and three varieties used in the preparation of white beiju (Bonita, Branquinha, and Castanha).
Variation and Local Classification of Soils Managed by the Communities Farmers classify soils mainly by their color, texture, and suitability for agriculture. Terras preta soils are considered very suitable for cultivation, and in early stages of succession have a dense cover of shrubs and vines. These soils also have indicator plant species (according to farmers’ perceptions), such as mucajá (Acrocomia aculeata (Jacq.) Lodd. ex Mart.), taperebá (Spondias mombin L.), paricá (Schizolobium amazonicum Ducke), samaúma (Ceiba pentandra (L.) Gaertn.). Terras preta soils also have a high abundance of fire ants (Solenopsis sp.). Local residents always associate these soils with the presence of ceramic fragments. Barro curuba, barro preto, barro solto and areia preta soils are considered to be of intermediate suitability for cultivation and have less dark coloration and fewer ants when compared to terras pretas. These soils are generally located close to small streams and ceramic fragments are occasionaly found in them. Chemically these soils are very heterogeneous and resemble anthropogenic soils due to high levels of OM and P; however, they are more acidic, showing higher levels of Al and less calcium and magnesium. Due to these differences, we categorize them as Btransitional soils^ (Fig. 3). Due to their abundance in the landscape and proximity to communities, these soils comprise most of the swiddens. Barro amarelo, areião and areia soils are considered the least suitable for agriculture. Most of these correspond to nutrient-poor Oxisols and Ultisols. When these soils are predominantly clayey (barro), they are considered by local farmers to be more fertile than sandy soils. The soils of the swiddens in the communities are very heterogeneous in their physical and chemical characteristics (Fig. 3). The first two axes of the PCA explained 77% of the variation in the dataset. The variables pH, Ca, Mg, P, K, Mn
Terra preta Transitional soils Oxisols / Ultisols
Sand
Enseada do Amorim
pH H2O
Parauá 10
6 6 1
Zn
PCA2 (20.3%)
5
Ca Mg Mn P
K
5
Fe
N
8 Surucuá Fig. 2 Number of common and exclusive manioc varieties managed in 61 fields in three riverside communities in the lower Tapajós River, Pará, Brazil
OM Al Clay
PCA1 (56.8%)
Fig. 3 Principal Component Analysis (PCA) of chemical and physical variables of soils collected in 61 fields in the Resex Tapajós-Arapiuns, lower Tapajós River region, Pará, Brazil. Numbers in brackets show the percentage of variation of soil properties explained by each axis. Each point is a field and the vectors are the measured variables
Hum Ecol
and Zn were negatively correlated with the first PCA axis, while Fe was positively correlated. This axis can be interpreted as a fertility axis, representing the variation from the most fertile anthropogenic soils (located to the left of Fig. 3) to the adjacent low fertility soils (located to the right). The PCA2 axis is negatively correlated with clay, organic matter and Al, and positively with sand. This axis summarizes mostly the variation in soil texture. The fields located in terra preta have higher contents of P, Ca, Mg, Zn, Mn and organic matter, higher pH and lower levels of Fe and Al. In contrast, the fields in Btransitional^ soils have higher levels of Al, Fe and organic matter, lower levels of Ca, Mg, Mn, Zn and lower pH, although this is a rather heterogeneous category in terms of soil physical and chemical characteristics. The Oxisols and Ultisols have the lowest values of P, K, Ca, Mg, Zn, Mn and organic matter, and the highest levels of Fe and Al, lower pH and high concentration of sand.
Soils Influence the Choice of Manioc Varieties The assemblages of manioc varieties managed in the fields of smallholder farmers of the lower Tapajós River are related to soil type (Fig. 4). The variety BMulata^, grown by 50% of households, is mostly grown in the fields on terra preta, occupying almost 50% of the area cultivated in these soils. The time to maturity of this variety is six months and after 12 months the roots start to rot; this variety is categorized as weak or watery. In contrast, the varieties BAchadinha,^ grown by 67% of households, and BCuruá-una,^ grown by 77% of households, occupy larger areas in fields on Ultisols and Oxisols. Both varieties have a longer time to maturity (Table 1) and only start to rot after 24 months; they are
categorized as strong and/or dry (higher concentration of starch and higher flour yields). Together, these varieties occupy more than 60% of the area cultivated in less fertile soils. In addition, the varieties BPretinha-amarela^ (37% of households), and BAcarizinha^ (43% of households), are considered intermediate in the classification of strength (Table 1), with intermediate maturation times and time to rot (18 months). These five most common varieties occupy more or less equivalent areas in transitional soils (Fig. 5). The sweet manioc varieties are fast-maturing varieties used mainly for domestic consumption and are managed in all three soil types, but they occupy slightly larger areas in terra preta. These results show that farmers prefer to cultivate weak varieties on ADE and strong varieties on Oxisols and Ultisols. Farmers who grow manioc on transitional soils grow a mixture of fast, intermediate, and late maturing varieties in similar proportions (Fig. 5). Households who maintain swiddens in different types of soils manage different groups of varieties in each soil. Farmers also adjust the way they manage their fields depending on the type of soil. The duration of the fallow period and the number of successive cycles were significantly different (Fallow duration: F = 30.24; p < 0.001 and planting cycles: F = 17.78; p < 0.001) among the three soil types (Fig. 6). In anthropogenic soils the fallow duration is shorter and the number of successive cycles is larger (up to four cycles; Fig. 6). In Oxisols and Ultisols the fallow duration is longer and the fields have a maximum of two successive cycles. In transitional soils both fallow duration and the number of successive cycles are intermediate. For farmers, the preference for cultivating weak short-cycle varieties in ADE is explained as a strategy to avoid the high labour demands for cultivating these anthropogenic soils (mostly for weeding).
Anthropogenic soils Transitional soils Oxisols/Ultisols
Transitional soils
Terra preta
Mulata (0.2) 0.5
Acarizinha (0.3)
3
0.4
Curua-una (0.6)
0.3
11
2
0.2
0.1
Pretinha amarela (0.4)
7 0
0
Achadinha (0.7)
14 Macaxeira-boliviana
3 Oxisols / Ultisols Fig. 4 Common and exclusive manioc varieties grown in three soil types in three riverside communities in the Resex Tapajós-Arapiuns, lower Tapajós River, Pará, Brazil. Number of fields sampled: 8 terra preta; 37 transitional soils; 16 Oxisols/Ultisols
Macaxeira-branca Macaxeira-manteiga
Fig. 5 Manioc varieties that are most commonly cultivated by farmers of three communities in the lower Tapajós River (30 families and 61 fields) occupy different proportions of fields in different soil types. Numbers along the vertical axis and points indicate the proportion of the area of each soil category that is cultivated with each variety. Values were obtained from the sum of the area that the variety occupies in a given soil divided by the total area cultivated in this soil. Numbers after the name of the varieties are the Strength Index
Hum Ecol Fallow duration
5
Planting cycles
#a
Length of fallow (years)
10
4
A
8
3
B 6
b C
2
c
4
1
2
0
Number of successive cycles
12
0
Terra preta
Transitional
Oxisols / Ultisols
Fig. 6 Differences in the management of swiddens in different soil types along the lower Tapajós River, Pará, Brazil. Means estimated for terra preta soils (8 swiddens), transitional soils (37 swiddens), and Oxisols/ Ultisols (16 swiddens). Vertical bars show the standard deviation. Means with different letters differ significantly by the Tukey test at 5% probability
According to farmers, the number of consecutive cycles that can be grown in a given area is associated with the successional stage of the vegetation that was cut to establish the swidden. Usually, fields that are cut from fallows in early stages of succession (called ‘low capoeira’ or ‘weak capoeira’) are cultivated for fewer cycles than those cut from older fallows. In contrast, in fields cut from areas with more advanced successional stages, called ‘high capoeira’ or ‘strong capoeira’ (over 12 years of fallow), manioc can be planted for two consecutive cycles before fallowing. On anthropogenic soils, however, manioc can be planted for several consecutive cycles, even if the field had been cut from a low capoeira. Strong manioc varieties are considered by farmers to be suitable for cultivation in ‘strong land’, i.e., fields opened in high forest areas and/or in ‘mature capoeira,’ while weak ones are suitable for cultivation in ‘weak capoeira’ (short fallow), such as ADE.
Relationships between Market Orientation and the Diversity of Manioc Varieties The 30 households produced an average of 406.0 ± 220.7 kg of manioc flour (8 sacks of 50 kg) per month, but with ample variation among households (100–900 kg or 2–18 sacks). Production is also variable during the year; it is generally lower from October to March (the period when fallows are
cut and burned, swiddens are planted and manioc grows towards maturity) and higher from April to September (harvest season). The percentage of flour production that each household designated for the market ranged from 30 to 88.9%. The sale of flour occurs both in nearby urban centers, e.g., Santarém, and within and between communities. Intensely yellow flour with no added colorants and small and homogeneous grains has the highest market value. Twenty-seven of the 36 varieties used for flour production have yellow or yellowish pulp. Farmers consider that the yellow varieties often have less starch, are watery, compared to the cream and white varieties, which often have higher yields. To balance productivity and market demand, farmers mix the yellow varieties with whiter varieties during the production of flour. The number of varieties cultivated per household ranged from 3 to 9, with an average of 5.8 ± 1.5 varieties (Table 2) and was not influenced by market orientation (p = 0.64; Fig. 7a). Commercial varieties tend to occupy larger areas in the fields of households with a higher market orientation (p = 0.03; Fig. 7b). The size of the swiddens was positively related to market orientation (p < 0.001; Fig. 7c).
Discussion The number of manioc varieties found in this study is similar to the range reported in recent studies involving riverside communities in Amazonia. Fraser et al. (2012) reported 50 varieties for 249 households in six communities along the middle Madeira River; Lima et al. (2012) found 54 varieties grown by 55 households in 13 communities along the middle Solimões River; Robert et al. (2012) reported 47 varieties among 33 farmers in two Mebêngôkre-Kayapó indigenous communities, Xingu River; Lima et al. (2013) found 52 varieties among 46 farmers in eight communities in Santarém and Juruti, Pará. Taken together, this body of research shows that, despite increasing market orientation in the lower Tapajós River, the communities still maintain a considerable number of manioc varieties. However, market orientation may lead farmers to discard varieties that do not meet market demands and turn their attention to commercial varieties in greater demand. This
Table 2 The proportion of manioc flour directed to the market, the size of the cultivated area, the number of managed varieties and the proportion of the area planted with commercial varieties in the 61 swiddens of 30 households in the Resex Tapajós-Arapiuns, lower Tapajós River, Pará, Brazil. Values are means, followed by standard deviations Production for market (%)*
Cultivated area (ha)
N° varieties
Area in commercial var. (%)**
