Economic Botany

, Volume 63, Issue 3, pp 242–255

A Review of Perceptual Distinctiveness in Landraces Including an Analysis of How Its Roles Have Been Overlooked in Plant Breeding for Low-Input Farming Systems


    • Natural Resources InstituteUniversity of Greenwich

DOI: 10.1007/s12231-009-9086-3

Cite this article as:
Gibson, R.W. Econ Bot (2009) 63: 242. doi:10.1007/s12231-009-9086-3


A Review of Perceptual Distinctiveness in Landraces Including an Analysis of How Its Roles Have Been Overlooked in Plant Breeding for Low-Input Farming Systems. Traits providing perceptual distinctiveness (PD), which allow less commercial farmers in developing countries to recognize and name individual landraces, enable the creation and management of their diversity and the transfer of knowledge of each to other farmers and succeeding generations. Worldwide examples illustrate how PD traits on seeds and vegetative propagules help maintain genetic purity and provide markers at planting time, identifying landraces suitable for planting at particular locations and times and for future household and market needs. PD traits on the yield also enable household members and customers to identify and value landraces for different uses. To fulfill these roles, they are generally highly salient, restricted in number, environment-independent, qualitatively inherited, generally with expression based on one or a few genes, and often culturally significant. Even so, they are seldom mentioned as varietal selection criteria by farmers, who may be unaware of their importance, or in plant breeding programs and in situ conservation of plant genetic resources projects; the need for national variety release committees and policymakers in developing countries to include them is emphasized.

Key Words

In situ conservationDiversityClimate changeVarietal improvementAgromorphological traitTraditional varietySubsistence farmer


A diversity of phenotypically-distinguishable landraces of the same crop species are commonly grown by individuals and communities of less commercial farmers in low-input systems in the tropics and semitropics (to whom the term “farmer” refers in this review unless otherwise specified). Andean potato landraces provide an extreme example, some 5,000 being recognized, with >100 grown in a single valley and >12 grown by a household (Brush 1992). Other examples: in lowland South America, just 16 Amerindian communities grew >200 landraces of cassava (Salick et al. 1997), one community of 30 households alone growing 76 (Elias et al. 2001a); some 600 landraces of beans were reported in Rwanda (Sperling and Berkowitz 1994); and 103 and 89 rice landraces were grown at one ecosite in Vietnam (Tin 1999) and by the small Baduy community of West Java (Iskandar and Ellen 1999), respectively. Landraces of the same species may be cropped together: potato farmers in the Andean region have been reported to grow an average of 21.1 landraces/225 plants sampled/field (Zimmerer 1991), with one field containing 46 landraces (Brush et al. 1981), and, in central Africa, 42 Rwandan households have been reported to grow mixtures of 6–29 (averaging 20) bean landraces/field (Voss 1992). This diversity satisfies multiple needs (food, forage, brewing, cultural, religious, etc.) and markets, stabilizes/increases productivity in diverse and unpredictable environments, and helps resist attack by pests, weeds, and diseases (IPGRI 2001; Jarvis et al. 2006; Smithson and Lenné 1996).

The single most important management tool of farmers for maintaining their diverse landraces in complex cropping systems is perceptually distinctive morphological traits that enable each landrace to be distinguished (Boster 1984, 1985; Clawson 1985). Despite this, reviews of even participatory plant breeding (PPB) and participatory varietal selection (PVS), which primarily target poorer farmers in developing countries (Almekinders and Elings 2001; Vernooy 2003; Weltzien et al. 2003), give little prominence to perceptual distinctiveness, consistent with personal experiences of farmers also not mentioning it during cassava (Manu-Aduening et al. 2006) and sweet potato (Gibson et al. 2008) PPB. This identified the need to increase awareness of:
  • ➢ the importance, roles, and properties of traits providing perceptual distinctiveness among landraces;

  • ➢ their potential roles in plant breeding, especially PPB; and

  • ➢ their consideration in breeding and in situ conservation of crop diversity policies.

Some Terms Used in This Review

In this review, “variety” is used for an officially released cultivar and “landrace” for a cultivar developed and maintained by farmers using traditional methods; a landrace is a perception of farmers and may be genetically heterogeneous (Ceccarelli and Grando 1991). The term “cultivar” is used in the broad sense to describe any defined cultivated plant population; it therefore includes both varieties and landraces and is used if it is unclear which term is appropriate or if neither may be, for example, for the product of PPB. Traits with the apparent main or sole purpose of providing perceptual distinctiveness (PD; also used as an abbreviation for perceptually distinctive) (Boster 1984, 1985) are distinguished from “use traits,” which directly make a cultivar fit for the human purpose of yielding the quality and quantity of required products in different agroecological or technological circumstances (Bellon 1997).

The Importance of Perceptual Distinctiveness between Landraces

Across crop species, continents, and human cultures, landraces are renowned for their diverse appearance, signifying the special importance of PD traits to man (Darwin 1859). This is particularly evident in landrace seeds (Zeven 2000), their color and/or patterning contrasting sharply with the general uniformity and monochrome of wild plant seeds, including those of crop progenitors. In particular, reviews of maize landraces in the Americas (Clawson 1985; Hernandez Xolocotzi 1985; Johannessen et al. 1970; Zimmerer 1991) and subsequent literature (Benz et al. 2007; Gomez et al. 2000; Louette 1991; Perales et al. 2003; Steinberg 1999) describe a wide range of seed colors (white, yellow, blue, pink, red, black, gray-blue uniform coloration x speckling [pinto]). Examples from other crops and cultural backgrounds include the myriad colors and color patterns on seeds of common bean in Malawi (Martin and Adams 1987) and on tubers of Andean potato landraces ( (Brush et al. 1981), Hopi native Americans growing 17 maize, 4 lima bean, 3 string bean, 7 field bean, and 2 tepary bean landraces, each with different colored seeds (Soleri and Cleveland 1993), and Baduy people living in upland West Java distinguishing 89 rice landraces from the amount, color, length, and form of the hairs of the grain and the color of the husked grains (Iskandar and Ellen 1999). Further indicators of the importance of PD traits to traditional farmers are the following:
  • ➢ In a review of 30 studies of maize domestication by 25 Amerindian groups spanning 1750–1963, there was a similar number of mentions of seed color (21) and of use traits (27) (Johannessen et al. 1970) and, in studies on ensete (Shigeta 1996) and common bean (Mekbib 1999), farmers ranked PD traits above use traits.

  • ➢ The importance of PD traits in folk classification of plants including landraces (Berlin 1976; Hays 1979); for example, the primacy of tuber shape and color in Andean potato (Brush et al. 1981; Zimmerer 1991), and of seed color in maize (Zimmerer 1991) and bean (Martin and Adams 1987) landrace classification. Names of landrace that are long established in a community are more likely to refer to PD traits than those of recently acquired landraces (Nuijten and Almekinder 2008).

  • ➢ PD traits such as cassava leaf shape and stem and petiole color (Boster 1985) and seed colors in maize (Clawson 1985), sorghum (Ayana and Bekele 1998; Geleta and Labuschagme 2005), and finger millet (Tsehaye et al. 2006) are culturally and geographically widespread.

  • ➢ Farmers in more subsistence-oriented communities in Mexico and the Philippines have maize landraces with more diverse seed colors (Gomez et al. 2000) and more subsistence-oriented farmers in the Philippines give greater priority to morphological traits of their sweet potato landraces (Nazarea 1998).

  • ➢ At least one gene providing pod and one providing seed color in bean are co-located with those for other traits recognized as part of the domestication syndrome (Koinange et al. 1996).

Evidence of the efficacy of PD is that farmers are more or less consistent in naming phenotypically (Boster 1984; Sambatti et al. 2001; Tunstall et al. 2001; Zimmerer 1991) and genetically (Briand et al. 1998; Brush et al. 1981; Mkumbira et al. 2003; Quiros et al. 1990; Vargas-Ponce et al. 2007) similar landraces, though consistency of naming may be restricted to extended families and communities and vary with agroecology (Nuijten and Almekinder 2008; Salick et al. 1997). Conversely, when genotypes are indistinct, farmers fail to maintain them as separate landraces; examples include some polyclonal cassava (Elias et al. 2000, 2001a and b) and potato (Quiros et al. 1990) landraces and a bean landrace of a visually uniform seed type but comprising two populations (Briand et al. 1998).

Roles of PD Traits

Resource-poor farmers commonly have multiple uses of what they have available and, in addition to enabling farmers to recognize and distinguish landraces, other reported roles of PD traits are:
  • ➢ Indicating particular qualities in the yield(s)

  • ➢ Identifying landraces for different agroecological circumstances and technological requirements

  • ➢ Maintaining genetic purity

  • ➢ Uses associated with specific rituals and beliefs

  • ➢ Recognition, adoption, and spread of new landraces

Indicating Particular Qualities in the Yield(s)

A crop species often has several yields—for example, maize yields grain, foliage, and a woody stem—and each yield may also have several uses. For example, grain may be sold or used by the farming household for food, livestock feed, and malt for brewing. PD traits on each yield of a landrace can indicate suitability for each use to household members or customers, in the latter case acting like a logo or trademark for a landrace during marketing. PD traits on seed also enable farmers to plan ahead by indicating its genetic fitness for producing a yield suitable for future uses or markets. If the seed or vegetative propagule is also the main yield, the same PD traits can fulfill both roles.

As examples, farmers plant black-seeded barley to provide future winter grazing (Ceccarelli and Grando 1991), and black-seeded finger millet is renowned for its straw quality (Tsehaye et al. 2006); white-seeded barley (Ceccarelli and Grando 1991), sorghum (Bernaud et al. 2006), and rice (Sthapit et al. 1996) are considered superior for eating. Finger millet landraces with different seed colors differ in resistance to birds, ease of threshing, storability, suitability for brewing beer and preparing different foods, and in market demand (Tsehaye et al. 2006). Maize with a yellow, pink, dark purple, or red grain are considered superior for animal feed, sweetcorn, tortillas, or pozolero, respectively (Hernandez Xolocotzi 1985), and their flours differ in cooking quality (Cabrera et al. 2001). In potato, PD traits by which Andean farmers distinguish their landraces, such as tuber shape and skin color and the number and shape of the eyes (Brush et al. 1981), are also market-related attributes (Thiele et al. 1997) recognized by farmers to act as “trademarks” (Potts et al. 1992). In cassava, the main yield is a root hidden underground, and foliar PD traits identify the appropriate landrace for harvest (Sambatti et al. 2001) and those that need processing to detoxify cyanogenic compounds (Mkumbira et al. 2003).

Most PD traits are indirect indicators of current or future usefulness, but some directly indicate suitability for particular uses. Thus, the characteristic depression of so-called dent maize is caused by the shrinkage of loosely-packed starch grains, and the kernels are therefore soft-grained and easily milled into flour, for example, for tortillas (Louette 1991); rice with translucent or opaque grains give nonglutinous or glutinous rice, respectively, when cooked (Iskandar and Ellen 1999; Lambert 1985). PD traits may also contribute aesthetic values; for example, purple-grained maize may be preferred for tamales (Hernandez Xolocotzi 1985) and provide the attractive red color of the maize infusion chicha morada. Such color traits may also provide health benefits: color of maize grain traditionally indicates medicinal properties (Hernandez Xolocotzi 1985), black- or blue-grained maize being considered by the Mopan Maya (Steinberg 1999) and Tzetzal (Benz et al. 2007), respectively, to be particularly strengthening, perhaps based on the antioxidant properties of anthocyanins (Del Pozo-Insfran et al. 2006). Flavanoid pigments of pigmented bean seeds may also provide health benefits (Hosfield 2001); darker sorghum grain also has more anthocyanins and other flavanoid antioxidants (Dykes and Rooney 2006).

Identifying Landraces for Different Agroecological Circumstances and Technological Requirements

Instead of using high inputs to provide uniform near-optimum growing conditions, less commercial farmers deploy their portfolio of landraces to achieve adequate yields:
  • ➢ Many landraces, each narrowly adapted to a particular set of ecological conditions, are more productive than a few broadly-adapted ones (Ceccarelli 1994).

  • ➢ Polycultivar cropping reduces pest and disease attack and provides yield stability (Smithson and Lenné 1996).

  • ➢ Landraces with different times to maturity
    • stabilize yield in inclement seasons and allow replanting if an initial sowing fails (Clawson 1985), and

    • allow the planting season to be staggered and the cropping season to be extended (Defoer et al. 1997; Louette et al. 1997) to allow labor to be used more efficiently and market and home needs to be addressed over a longer season.

To do the above, farmers use PD traits to recognize seed of each landrace, and Clawson (1985), in a review of staple tropical food crops, stated that “the cultivation of multiple varieties of staple food crops of varying color and maturation period [is] the most fundamental function of intra-specific diversity in small-scale tropical agriculture.” Consistent with this, PD traits are mostly located on the propagule irrespective of whether this is a seed as in maize (Hernandez Xolocotzi 1985; Johannessen et al. 1970), a tuber as in potato (Brush et al. 1981; Zimmerer 1991), or a cutting even when these are not the main yield, as in cassava (Boster 1985) propagated by mature stem cutting, ensete (Shigeta 1996) propagated by offsets, and sweet potato (Hays 1979) propagated by foliar cuttings. Furthermore, PD traits on propagules are associated by farmers with different agronomic qualities. Maize (Cabrera et al. 2001; Hernandez Xolocotzi 1985; Louette et al. 1997) and sorghum (Ayana and Bekele 1998) landraces with different maturity dates have differently-colored seed. Farmers in Uttar Pradesh know the different soil, water, and manure requirements of each of their landraces of wheat, finger millet, barnyard millet, soybean, and rice, and each one’s seed is distinctive (Tiwari and Das 1997). Farmers in the Americas know that maize landraces distinguishable by seed color have different rainfall (Soleri and Cleveland 1993), topographical (Benz et al. 2007; Rice et al. 1998), and soil requirements (Louette et al. 1997), as well as different maturity dates. In Ethiopia, early maturity of a white-seeded finger millet (cv. Tsa’ada) is exploited when there is a need to escape drought or to replace failed crops (Tsehaye et al. 2006). Rice landraces with different seed colors differ in drought resistance (Vaughan and Chang 1992), and Malaysian farmers, knowing the moisture and depth of water requirements of their landraces, plant them appropriately along a gradient around and into ponds and small streams (Lambert 1985). The complex mixtures of landraces of common bean traditionally planted in Africa (Martin and Adams 1987) result in part from farmers selecting different compositions of phenotypically-distinct seeds for different soils and crop combinations (Voss 1992).

Maintaining Genetic Purity

Although easy for clonally-propagated or predominantly self-pollinating crops, maintaining separate landraces of out-crossing seed-propagated species such as maize or sorghum is potentially difficult, especially for traditional farmers lacking knowledge of pollination and genetics (CIMMYT 2000; Manu-Aduening et al. 2005). Farmers can potentially minimize cross-pollination by planting maize landraces at a distance from one another or at different times (Halsey et al. 2005; Zeven 2000), but neither technique may be practiced (Bellon and Brush 1994). Instead, maize farmers in the Americas maintain landrace purity by rejecting cobs in which the grain color type of that landrace is atypical (Hernandez Xolocotzi 1985; Louette 1991; Steinberg 1999; Zimmerer 1991). This utilizes the xenia effect, pollination by another cultivar differing in a dominant seed endosperm trait such as color being expressed in the resulting seed, a phenomenon also used by researchers to map the spread of maize pollen (Halsey et al. 2005 and references therein). Rejection by farmers is, however, because of tradition; thus, cobs of a white maize landrace with some red grains are considered by Mopan Mayans in southern Belize to be splattered with Christ’s blood (Steinberg 1999)—and, because such seeds are likely to have been pollinated by another landrace, this rejection excludes contaminated seed (Louette and Smale 2000). Experimental use of the xenia effect to eliminate progenies of cross-pollination in artificially-intercropped maize varieties also selected for nonoverlapping flowering and pollen incompatibility and so reduced cross-pollination in succeeding generations (Paterniani 1969). In established landraces, selection for uniform grain size and number of rows of grain on a cob stabilized maturity date because seed size and row number are associated with length of growing period (Louette and Smale 2000).

Uses Associated with Specific Rituals and Beliefs

Staple food crops have a religious significance in many traditional societies (Cabrera et al. 2001) and this can provide a basis for maintaining landraces (Steinberg 1999), especially those with unusual traits. Cassava is central to the beliefs of the Amuesha of northern Peru, the shaman maintaining and selecting different landraces (Salick et al. 1997). Maize has a similar role for other Amerindian cultures: for Mopan Maya, maize landraces with different colored grains were created by gods (Steinberg 1999), various maize landraces are an essential part of Hopi ceremonial life (Soleri and Cleveland 1993), and, in Mexico, some red-grained maize is traditionally included in crops predominantly of other maize cultivars to protect the latter (Clawson 1985). Some predominantly rice-growing Asian cultures similarly have a Rice Goddess, distinctive landraces being grown in sacred swiddens (Iskandar and Ellen 1999) and specifically-colored grains being used in rituals (Lambert 1985).

Recognition, Adoption, and Spread of New Landraces

Man “delights in bright colors and curious and unusual variants” (Harlan 1975), and Boster (1985) argued that Aguaruna cassava farmers initially select for PD in its own right, an argument supported from studies with other Amerindian cassava farmers (Sambatti et al. 2001), Andean potato farmers (Zimmerer 1991), and Ethiopian ensete farmers (Shigeta 1996). Recorded examples of this leading to acquisition of a new landrace include:
  • Mexican farmers growing yellow maize after spotting the unusual seed on a path or on the stems of plants blown into trees (Rice et al. 1998)

  • Pahang Malay farmers developing the more robust, larger-grained, and heavier-yielding Gangsa puteh ekar musang landrace with a distinctive black-husked vitreous grain from off-types noticed among seed of rice Gangsa puteh (Lambert 1985)

  • Ethiopian farmers developing three landraces of wheat from seeds with unusual appearances found in the field and marketplace (Haddi and Berg 2006)

However, such interest leading to the development of new landraces may be restricted, for example, mainly to women farmers (Boster 1984; Nazarea 1998; Salick et al 1997), the shaman (Salick et al. 1997), or to particularly curious farmers (Johnson 1972; Nazarea 2005; Rhoades and Bebbington 1995). Of 200 Ghanaian cassava farmers interviewed, only 70 had ever used cuttings from cassava seedlings: 5 of these did it out of curiosity, 31 because they perceived valuable use traits in the selected seedlings, and 34 because at that time they lacked other sources of planting material (Manu-Aduening et al. 2005); in a Tanzanian community, only one farmer was experimenting with cassava seedlings (de Waal et al. 1997).

PD is also required for a cultivar to be maintained (Boster 1985), and “Cultivars that demand careful attention to be discriminated will not diffuse as rapidly, all things being equal, as those which are more readily distinguished” (Boster 1984), farmers needing to identify landraces to transfer information about them (Boster 1986). From experiences with Syrian farmers recognizing that black-seeded barley landraces are better adapted to dry areas and provide better winter feed than white-seeded ones, Ceccarelli and Grando (1991) proposed developing cultivars with specific seed colors to speed up adoption. It is not clear if they followed this through, but pigeon pea cv. NPP 670 was adopted rapidly by Kenyan farmers mainly because its seed and the growing plant were easily distinguished from landraces (Jones et al. 2001). The different phenotypes, including the seed and cob, of modern maize varieties from landraces also facilitated their adoption by Mexican farmers (Louette et al. 1997).

Properties of Traits Providing Perceptual Distinctiveness

In order to fulfill the above roles, PD traits generally are:
  • Salient (Boster 1985; Clawson 1985): Most comprise bright colors or patterns of colors (see previous sections) or easily-observed distinctive features such as clawed maize seed (Perales et al. 2003) and the shape of a leaf (Boster 1985) or a potato tuber and its eyes (Brush et al. 1981).

  • Few (Berlin 1976; Hays 1979): For example, key traits in cassava may be just petiole color, leaf shape, and stem color (Boster 1985), providing simplicity to the recognition system.

  • Unaffected by environment, providing consistency to recognition: For example, petiole color on cassava landraces remains true across environments, unlike most use traits (Elias et al. 2001a).

  • Not adversely affecting use traits and randomly linked to them (Boster 1984, 1985).

  • Qualitative inherited traits controlled by one or a few genes as in the flower, pod, and seed color of Mendel’s peas (

PD traits possessing all these properties are likely to be rare, and some not possessing all properties fully are used. Zeven (2000) identifies some adverse agronomic and other properties associated with particular seed colors; for example, tannins in brown-seeded sorghum are associated with prolonged seed dormancy, less digestible foliage for cattle, and less palatable grain for humans (Ayana and Bekele 1998; Reed et al. accessed 2007). Rarity also explains why the same PD trait may sometimes be associated (but randomly) with different use traits in different, geographically-isolated landraces (Boster 1985). Some PD traits such as maize (Coe 2001) and bean (McClean et al. 2002) seed color are controlled at several loci. However, selection for a single color of seed in a landrace, repeated genetic bottlenecks such as are generated by self-pollination, as in beans, or traditional farmer selection (Benz et al. 2007), such as selection of maize seed from only a few plants (Gibson et al. 2005), ensure that most such loci become homozygous. Consequently, only one or a few of the genes controlling a PD trait are likely to differ between any pair of landraces, and the trait is inherited according to Mendelian laws when crossed. More complex, polygenically-controlled differences, for example, in potato tuber size and shape (Brush et al. 1981) or in maize plant morphology, length, row number, and thickness of the cob (Sanchez et al. 2000), distinguish major races of landraces. This also allows landraces belonging to different races to share PD traits, for example, the same seed color (Arias et al. 1999; Louette 1991; Perales et al. 2005).

Apparent Lack of Purposeful Use of PD Traits in Plant Breeding

Traditional Farmer Breeding

The previous section on recognition, adoption, and spread of new cultivars reports how farmers select unusual plant variants, and intellectual interest in diversity of PD per se may to an extent be responsible for landrace diversity. However, once a PD trait has been noticed by farmers within a population of plants, any association with a use, for example, yielding well on poor soil or early maturity, is tightened by farmers consistently planting it on poor soil or late in the growing season, only genotypes that do well under such circumstances generating much, if any, seed (Louette and Smale 2000). This process over time generates a portfolio of PD landraces for different uses (Boster 1985). While interest in unusual variants is clearly conscious, farmers may be unaware of it as a first step in landrace development (Boster 1984, 1985; Shigeta 1996), being more concerned with use traits when asked about their varietal needs (Cleveland et al. 2000). Examples include Ethiopian farmers of sorghum (Teshome et al. 1999; Tunstall et al. 2001), Malian (men and women) (Defoer et al. 1997), and Mexican farmers of maize (Perales et al. 2003) and Indian farmers of pearl millet (Choudhary et al. 1997). Similar results have been obtained when farmer criteria have been sought as part of PPB (see next section).

Participatory Plant Breeding (PPB)

PPB and in situ conservation of crop diversity (IPGRI 2001; Jarvis et al. 2006), which share common ground (Almekinders and Elings 2001), target cultivars grown by less commercial farmers (Vernooy 2003; Weltzien et al. 2003), often enhancing participation and empowerment of poorer and women farmers (Sperling and Berkowitz 1994). They may include studies of how farmers manage crop diversity (Louette and Smale 2000; Rice et al. 1998; Soleri et al. 2000), aiming to incorporate traditional knowledge of breeding (Sperling and Berkowitz 1994) or addressing weaknesses through sharing scientific knowledge (Almekinders and Elings 2001; Gibson et al. 2005, 2008; Louette and Smale 2000; Manu-Aduening et al. 2006; Soleri et al. 2000). PD itself, however, has received little recognition: examples include PPB of beans in Rwanda (Sperling and Berkowitz 1994), pearl millet in Niger (Baidu-Forson 1997), Burkina Faso, Mali, Nigeria (Omanya et al. 2007), and Rajasthan (Weltzien et al. 1996), rice in India (Virk et al. 2003) and in Nepal (Sthapit et al. 1996; Joshi et al. 1997), barley in North Africa (Ceccarelli et al. 2001), cassava in Ghana (Manu-Aduening et al. 2006), sweet potato in Uganda (Gibson et al. 2008), and various vegetables in southeast Asia (Smolders 2006). The few PD traits that were mentioned mainly involved post-harvest attributes recognized by farmers for their use in marketing (Humphries et al. 2005; Kornegay et al. 1996; Mekbib 1997; Potts et al. 1992; Sthapit et al. 1996). A few PPB studies initially identified PD traits but only evaluated use traits (Mekbib 1999; Soleri et al. 2000; Thiele et al. 1997). Honduran but not Mexican farmers included maize grain color as a selection criterion (Smith et al. 2001); although Mexican farmers stipulated PD traits to be required for an improved maize population to be typical (“legitimo”) of the original landrace (Louette 1991), they specified use traits (large cob well filled with healthy seeds) as seed selection criteria for PPB (Louette and Smale 2000). Only use and market-related PD traits were considered important in the rapid spread of a PVS-selected rice cultivar (Witcombe et al. 1999).

Formal Plant Breeding: DUS and VCU Requirements for Variety Release

Official release usually requires that a variety is distinct, uniform, and stable (DUS) and has value for cultivation and use (VCU), criteria increasingly being adopted in developing countries (Anonymous 2005). DUS criteria resemble PD traits only superficially: though both provide identification, the former may be at any stage of a crop and are used for variety authentication, the latter are stage specific and use associated. Despite previous pleas (Clawson 1985), PD traits continue to be overlooked by formal plant breeding, and seeds of most modern varieties targeting developing country farmers are not distinct; for example, most maize varieties in Africa have white grains and have overwhelmed the previous diversity of colors (McCann 2005). Even when PD traits were identified as the primary reason for rapid adoption of a modern pigeon pea variety in Kenya (“Four varietal characteristics favoured diffusion of this seed. First, NPP 670 is easily distinguished from local pigeon pea varieties because of its determinate growth habit, short-stature and bold white seeds.”), their value was not among lessons mentioned as learned (Jones et al. 2001).


This review collates reports from a range of disciplines and cultural, geographical, and crop sources to develop a comprehensive framework for the roles and properties of PD traits (Fig. 1). In doing so, it has confirmed PD as the primary tool by which the diversity of landraces still characteristic of less commercial farming in developing countries was created and then maintained through succeeding generations. PD traits occurring on the propagule enable farmers to plant them as required to fulfill future demands, and in appropriate agronomic conditions, to maintain their genetic purity and to facilitate planting material exchange. PD traits on the yield enable landrace products to be prepared and used appropriately, including for marketing and religious festivals. This wide range of roles helps explain why such a phenotypic diversity of landraces is required. PD traits also allow farmers to name landraces and so exchange information about them to both succeeding generations and other communities. This holistic view of PD traits explains their key properties (highly salient, few, environment independent, qualitatively inherited, without adverse effects on yield, and randomly linked to use traits), how these enable PD traits to fulfill their roles, and the potential benefits of breeding PD traits into new cultivars, especially for less commercial farmers.
Fig. 1

The main properties, roles, and potential benefits of incorporating PD traits into breeding programs

Landraces initially develop because someone spots a PD variant and maintains it, perhaps driven by their cultural position, circumstances, or curiosity or because it also appears to have superior use traits. Selection and use during succeeding crop generations may associate the PD trait with particular crop use traits; this may be more or less random, farmers appearing generally unaware of how PD traits (other than those involved in marketing) help them succeed in their farming activities (Boster 1984, 1985). Once established, their use and maintenance arguably occurs within habitual activities, farmers seldom mentioning them as important varietal criteria (see section Participatory Plant Breeding). PD traits may, however, seem too obvious to farmers to mention to researchers, requiring careful questioning (Cleveland, pers. comm.) to elicit their importance (Johannessen et al. 1970; Mekbib 1999; Nazarea 1998; Shigeta 1996). A parallel may also be seen with traditional classifications of plants and animal, particularly the debate between utilitarian and intellectualist approaches (Berlin 1992). Clearly PD traits are useful to farmers; however, the fact that farmers and shaman associate some PD traits of landraces with beliefs and ceremonies (see Uses Associated with Specific Rituals and Beliefs) is consistent with findings in traditional taxonomy that beliefs alter viewpoints (“Why Is the Cassowary Not a Bird?”: Bulmer 1967) and may be a starting point for further research on farmer awareness of PD traits and of their roles.

Although PD traits are widely used by plant taxonomists to distinguish landraces, most plant breeders view landraces as sources of diversity for developing new varieties with improved economic traits, and PD traits and other significant aspects of farmers’ selection practices (Cleveland and Soleri 2007) have been overlooked even in PPB (see Participatory Plant Breeding). A major purpose of this review is therefore to make particularly scientists involved in PPB and in situ conservation of crop diversity more aware of the several roles of PD traits and their potential to deliver major benefits, especially in the sustained adoption and dissemination of new cultivars by communities. A re-emphasis of ethnobotany as declared in the Kaua’i Declaration (Anonymous 2007) would also be a major step towards achieving this. The threat to agriculture of climate change (Parry et al. 2007), to which poorer families and crop yields in tropical regions are particularly vulnerable (Schneider et al. 2007), make this re-emphasis particularly important. PD traits provide the means by which vulnerable farmers can use the diversity of their landraces to ameliorate effects on crop production (IPGRI 2001; Jarvis et al. 2006; Smithson and Lenné 1996) against predicted increases in temperature and changes in rainfall, increased variability and increased extremes of weather, including in Africa (Boko et al. 2007).

As well as benefits to be gained by incorporating PD traits into plant breeding, costs of continued ignoring of PD traits in PPB and conservation of crop diversity projects and publicly-funded on-station breeding programs for low-input farming systems in developing countries are considerable. Absence of suitable PD traits on planting material of new varieties thwarts traditional methods of seed maintenance, witting or unwitting overlap with PD traits in landraces causing the latter to be lost or, more likely, become mongrels. Mimicry of PD traits on the yield confuses customers into buying an imitation product (Thiele et al. 1997), potentially disappointing them and so destroying markets for established landraces. Mimicry of PD traits is also a form of infringement of brands/trademarks and associated intellectual property rights and needs to be controlled for ethical reasons. National variety release committees and policymakers in developing countries in particular need to be made aware of the importance and manifold roles of PD traits and the costs of ignoring them.


Richard Lamboll at NRI and David A. Cleveland at the University of California, Santa Barbara, and two unnamed reviewers provided valuable criticism.

Copyright information

© The New York Botanical Garden 2009