Diversity of Common Bean (Phaseolus vulgaris L.) and Runner Bean (Phaseolus coccineus L.) Landraces in Rural Communities in the Andes Highlands of Cotacachi—Ecuador

Abstract

Phaseolus species are cultivated worldwide as a primary food source for human consumption. Common bean (Phaseolus vulgaris L.) and runner bean (Phaseolus coccineus L.) landraces are often cultivated together. The purpose was to document the traditional knowledge held by farmers regarding bean landraces, to describe the diversity through seed morphological descriptors (five quantitative and five qualitative traits), to understand the geographic distribution using Moran’s I statistic, and to analyze the diversity through the Shannon–Wiener Diversity–Equity (H′) index and the Shannon J’ homogeneity index at the community level to better understand the richness of this geographic area. A total of 361 common and runner bean accessions were collected in 10 Andean rural communities of Cotacachi, North Ecuador. We identified 47 landraces, predominantly kept by indigenous female farmers, with limited access to formal education, highlighting the role of this disadvantaged population in agrobiodiversity conservation. The cluster analysis revealed three groups (cophenetic correlation coefficient = 0.6). In the principal component (PC) analysis, 94% of the variation was explained by PC1 and PC2. A positive spatial autocorrelation (Moran’s I: 0.24; z-score: 2.20; p-value: 0.03) was identified, confirming a spatial structure. The Morochos community showed the highest diversity (H′ = 1.55). The information of the diversity and distribution of common and runner bean variability provided in this study is a contribution to further research focused on conservation.

1. Introduction

The Fabaceae family is an important contributor to food systems [1], and it provides 33% of the human dietary protein consumption [2,3]. It has about 770 genera and 19,500 species [4,5]. There are genera within the family that are important food sources for human feeding. The genera Phaseolus, Pisum, Vicia, Cicer, Glycine, Lens, Cajanus, Vigna, Arachis, and Lupinus are considered an alternative to animal protein and can improve the daily diet of families in most areas of the world [3,5,6,7]. The genus Phaseolus includes 75 infrageneric taxa [8]. There are species within the Phaseolus genus such as P. vulgaris, P. lunatus, P. coccineus, P. dumosus, and P. acutifolius that are cultivated around the world as a source of food for human consumption [9,10,11,12]. Common bean (Phaseolus vulgaris L.) and runner bean (Phaseolus coccineus L.) are usually cultivated together [13]. These two species are very closely related and are partially inter-crossable when P. vulgaris is the female parent. Common bean is an annual plant predominantly autogamous, while runner bean is mainly allogamous and considered a perennial plant [13,14].

Common bean is the most planted legume in the world [1,15]. It is estimated that the area harvested in 2022 was about 36,792,490 ha [16]. It involves wild and domesticated forms [17]. It has a significant nutritional value and health properties [14,18,19,20], as well as socioeconomic importance [21,22,23]. Furthermore, it is one of the most studied and ex situ conserved species worldwide [24], as well as one of the most cultivated species in traditional farming [25].

There have been very few studies into the origins of P. coccineus. As mentioned by Bitocchi et al. [14], two studies have addressed the domestication of this species. On the other hand, research on the history of P. vulgaris suggests that there was a divergence in wild bean populations approximately 165,000 years ago, originating from a Mesoamerican ancestral population [26,27,28]. Common bean was domesticated after this divergence [29] about 8000 years ago [27,30] independently in two main geographic locations: Mesoamerica and the Andes, deriving two major eco-geographically distinct domesticated gene pools [22,27,31,32,33,34,35,36]. Nevertheless, important gene exchange between these two gene pools and genetic material movement have occurred since domestication [15]. In general, Mesoamerican genotypes show small- or medium-sized seeds, while the Andean genotypes are typically larger [37]. Due to the process of domestication, a wide number of varieties have been generated [18,35].

Traditional varieties, known as landraces or farmers’ varieties, are an important element of agrobiodiversity [1]. Crop landraces are local varieties with agricultural, cultural, and historical value, usually genetically diverse [38,39], that have been adapted to a specific ecosystem through a selection process developed by farmers [15]. Landrace diversity is important and should be considered as a supply of genetic resources for crop breeding [1]. Farmers usually give these landraces different names based on criteria such as growth habit, phenology, seed morphology, and use [40].

In this context, traditional-based agriculture (small-scale production) is an important source of food from the perspective of cultural identity and values. Usually managed by households [41], it allows to conserve crop genetic diversity, as well as to support ecosystem functioning, resilience, and productivity [42,43]. The conservation of crop genetic resources is important for local and global food security, as well as to face population growth and global warming [44,45]. That is why landrace in situ conservation should be considered to improve the agroecosystem structure [43].

Unfortunately, the world is facing a considerable rate of agrobiodiversity loss [38,46,47]. There is an increasing reduction in genetic variability due to intensive anthropic interventions [8], the progressive substitution of landraces with high-yield modern varieties by farming communities [43], land use change, and habitat degradation [16,44], as well as cultural transformation and diet changes [23,48].

That is why understanding the dynamics of plant diversity within a specific area and its cultural environment is basic for the conservation of agrobiodiversity. By increasing the knowledge of these factors, we can better appreciate the importance of preserving the unique diversity and cultural heritage of a region. Consequently, inventories play an important role in obtaining information about farmers’ varieties, which can then be used to promote conservation and reduce the risk of genetic erosion [1,17]. Even when a genetic diversity assessment is a fundamental principle of crop breeding programs and plays an important role in germplasm conservation [49], by documenting these varieties, we can better understand their diversity and characteristics, leading to improve conservation efforts and sustainable agricultural practices. Furthermore, the collection and characterization of bean landraces are important to prevent the loss of unknown genetic resources [50].

In this context, we decided to work on this research as a contribution to better understand the bean diversity of an area known for its traditional-based agriculture. The area we selected is Cotacachi, in the north of Ecuador, which is considered an important center of agrobiodiversity [51,52]. Being crossed by the Andes Mountain range, it has several altitudinal belts, which favor climatic diversification, responsible for the development of extensive agricultural biodiversity. The agricultural calendar is regulated by the rainy seasons, which are characterized by higher levels of rainfall, occurring in March and April and October and November, with average rainfall of 111 mm/month and 96 mm/month, respectively [53]. In contrast, the dry season takes place during the months of July and August, with an average precipitation of 16 mm/month [53], accompanied by intense solar radiation and strong winds [54]. Throughout the remainder of the year, the area experiences moderate rainfall. The climate is high Andean equatorial, with constant temperatures throughout the year. At 2500 m.a.s.l., the average annual temperature is around 13.5 °C, and the average annual rainfall is approximately 820 mm, with a notable scarcity of water for agricultural use in the dry season [53,55].

The main source of food and economic income in the study area is family farming [56,57]. Farming is conducted starting at approximately 2300 m.a.s.l. in the Inter-Andean Valley and extended upwards along the Cotacachi Volcano slopes, until reaching an elevation of approximately 3300 m.a.s.l. [58]. Their farms have a surface area of less than 1 ha [59]. This agriculture system is locally called chakra (a Kichwa term, Kichwa words written in italic), and it is defined as an ancestral agricultural system of the Kichwa indigenous peoples in the Ecuadorian Andean region, as well as in other areas of Perú and Bolivia [60]. It is a small plot next to the house displaying a wide range of agrobiodiversity where a variety of traditional agricultural practices are applied [61]. The traditional practices of seed selection and on-farm savings are still very common in these indigenous communities [62]. The water supply is usually limited; thus, about 55% of families do not have access to water for irrigation purposes [63].

Therefore, the aims of this research were to record farmers’ knowledge on bean landraces (P. vulgaris and P. coccineus) and to describe its diversity through morphological characterization of the seeds to provide a view of the available genetic resources of the species in Cotacachi as a contribution to better understanding the agrobiodiversity richness in the area.

2. Materials and Methods

2.1. Study Area

This research was conducted in three administrative zones: Imantag, El Sagrario, and Quiroga (Figure 1) in the Andes Highlands (altitude >2500 m.a.s.l.) of Cotacachi Canton. It is located in the north of Ecuador, approximately 100 km from the capital, Quito.

Cotacachi has one of the highest concentrations of indigenous populations (41%) in the country [64,65]. The population is mainly bilingual (Spanish and Kichwa) [66]. Most of the communities in the high Andean area are part of Union of Peasant and Indigenous Organizations of Cotacachi (UNORCAC), which is a second-tier, non-profit organization created in 1977. UNORCAC is made up of 41 communities and various indigenous organizations. Their members have strong cultural and territorial identity and usually revalue and preserve ancestral knowledge and traditional practices [63].

2.2. Sampling

We randomly selected 87 smallholder farmers (

Table 1. Number of smallholder farmers selected for the study in each of the ten communities in three administrative zones of Canton Cotacachi.

Community	Total Number of Smallholder Farmers	Selected Smallholder Farmers
El Cercado	146	14
San Pedro	88	9
Topo Grande	103	10
Iltaqui	36	4
Chilcapamba	72	7
Morochos	125	13
San Antonio del Punje	40	4
Cumbas Conde	119	11
Arrayanes	35	4
Morlán	108	11
Total	872	87

), which represent 10% of the total farmers from 10 communities (out of 41 communities that are part of UNORCAC). These communities were chosen based on their focus on family farming and its role as traditional agriculture management representative examples in the area. Communities not centered in agriculture were not considered in this study.

We divided the research into three parts. The first one involved farmers’ information on the diversity of beans, the second part included seed morphological characterization conducted on 361 bean accessions collected in the chakras of the selected farmers, and the third part was related to the geographical distribution of the landraces and calculation of diversity indices at the community level.

2.3. Data Collection

2.3.1. Farmers’ Information

We conducted semi-structured interviews with 87 smallholder farmers to obtain information about their agricultural diversity. This method has progressively been used to document farmers’ knowledge related to their crop landraces and traditional agricultural practices [67,68]. Before the interview, informed consent was obtained from each farmer. The interview included information about the names of the landraces as farmers know them, eliminating synonymy (different names referring to the same landrace) and homonymy (one name applied to two or more different landraces). We used farmer-named varieties, because it has been stablished that it reflects farmers’ own understanding of their crop diversity [69]. At the same time, we focused on the perception of farmers about the landraces they believe are in risk of disappearing. The interview had three sections, as shown in

Table 2. Sections in the semi-structured interview conducted with 87 smallholder farmers of ten communities in three administrative zones of Canton Cotacachi.

Section	Variable	Categories/Description
1. General information	Farmer’s name
	Age
	Gender	male, female
	Ethnic group	Indigenous, mestizo
	Language	Kichwa, Spanish, Kichwa and Spanish
	Education level	no formal education, primary (elementary), secondary, higher education
	Access to irrigation water	yes, no
	Land ownership	own, leased, community, borrowed
	Geographic coordinates	latitude, longitude, altitude
2. Bean diversity	Identification of landraces	Could you name the bean landraces do you grow in your chakra?
		Can we get samples of the varieties from your chakra?
		Can you identify the name of the landraces you have?
	Seed’s source	Inherited, market, public institution, seed exchange
	Production’s purpose	Self-consumption, market, self-consumption and market, seed production
3. Perception of risk	Landraces at risk of disappearing	What landraces do you think that are disappearing?
		What do you think that are the main reason?

.

The interviews took place in the chakras, and they were conducted mostly in Spanish. In a few cases, a Kichwa interpreter was required.

2.3.2. Morphological Characterization

Seed samples of bean landraces were collected in chakras from 10 communities (Table 1) between June 2019 and January 2020. Each sample had approximately 50 seeds, although, in some landraces, seeds were scarce (even less than 10 seeds/sample). The samples were sealed and labeled for further analysis. We collected a total of 361 accessions of different bean landraces.

In order to find intraspecific diversity, our study was focused on seed morphology through qualitative and quantitative descriptors, as other studies have shown that seed descriptors are the most discriminant for common bean [23,35,70,71,72]. Moreover, characterization has several purposes, including the provision of statistical data on the analyzed germplasm and the identification of any absent genotypes within a specific collection [73].

Qualitative Descriptors

We used five seed qualitative descriptors: primary color, secondary color, color around hilum, coat pattern, and shape.

Primary and secondary seed coat color was assessed in the dry grain by direct observation under natural daylight, establishing the primary and secondary colors using a modified scale from the International Board for Plant Genetic Resources [74]. Additionally, the hilum-surrounding color was defined, as it is considered an essential part of seed appearance [72]. To assess this trait, we used a modified scale [75].

The seed coat pattern was determined through the classification proposed by the International Board for Plant Genetic Resources [74]. To assess seed shape, we used a classification in three categories: oval-, cuboid-, and kidney-shaped, as applied in previous studies [71,76].

Quantitative Descriptors

We selected five seed quantitative descriptors previously used by Sinkovič et al. [35]. By using a digital caliper, we measured in millimeters (mm) 10 fully developed unharmed seeds in each sample in order to define the seed length (L) [highest parallel to hilum], seed width (W) [hilum to opposite side], seed thickness (T) [highest perpendicular to hilum], seed length/width ratio (L/W), and seed width/thickness ratio (W/T).

2.3.3. Geographic Distribution and Diversity Analysis

The geographic distribution was analyzed with software ArcGIS v. 10.8 [77], using Universal Transverse of Mercator (UTM) Projection, World Geodetic System (WGS) 1984 Datum, and 17 South Zone. Spatial autocorrelation Moran’s I statistic, with the Euclidean distance method, was performed to explain the spatial relation of the number of bean landraces among the chakras [78].

The Biodiversity Professional version 2 statistical analysis program [79] was employed to analyze the diversity of P. vulgaris and P. coccineus. Calculation of the agrobiodiversity indices was conducted at the community level using the Shannon–Wiener Diversity–Equity (H′) index, as well as the Shannon J’ homogeneity index.

2.4. Data Analysis

Statistical analyses were performed using the software InfoStat v. 2008 [80]. Statistical parameters for the quantitative descriptors included the mean, minimum and maximum values, standard deviation (SD), and coefficient of variation (CV) to assess the variability of the traits.

In addition, principal component and cluster analyses were executed [81]. Principal component analysis was, mainly, used to explore relationships among variables [82]. Cluster analysis was performed using the Ward method with Gower distance on standardized variables. It allowed to build a dendrogram, clustering the studied landraces into different groups according to their characteristics [69]. All the quantitative and qualitative traits of each accession were used. Analysis of variance (ANOVA) was performed, and the Fisher’s Least Significant Difference (LSD) method was used to compare means (α = 0.05).

3. Results

3.1. Farmers’ Information

The results show that 89% of the farmers identify themselves as indigenous, while the 11% remaining considered themselves as mestizo. Concerning the language, most farmers were bilingual (Kichwa and Spanish), although, in six cases, we needed an interpreter, because they only spoke Kichwa. Most of the farmers were female (63%), with an average age of 52 years old. In terms of education level, 41.4% of the farmers did not receive any formal education, 40.2% attended primary (elementary) school, 14.9% secondary education, and only 3.4% had access to higher education. Among the respondents, 4.6% rent the land they use to produce crops, while 8% use community land, and most of the farmers (87.4%) reported that they own their land. When considering the aspect of water accessibility, a significant majority of farmers (87.4%) mentioned that they do not have access to irrigation water and depend on weather conditions.

Regarding the perception of landraces at risk of disappearing, 26% of the farmers reported that Yana Vaca Poroto was the landrace they considered to be at risk of disappearing. However, we found this landrace in 10 of the 87 chakras. In contrast, we recorded two landraces (Poroto Campeon and Sucu Tomate Pintado) that were scarce in the study area, as they were found in only one chakra each.

3.2. Bean Landrace Diversity

Our study identified 47 landraces of bean species (Figure 2), from which 40 were common bean landraces (P. vulgaris) and 7 were runner bean landraces (P. coccineus). Common bean was usually planted in the middle of the chakras, while runner bean was planted along the borders. Locally, the generic name for common bean is poroto and, for runner bean, is popayan. As in previous research [40,83], we found that common names given to beans by local farmers in the Kichwa language are mainly based on seeds’ color and morphological characteristics. In our study, local names of the landraces seem to be related specially to the seed color, pattern, and shape, as described in other studies [84].

The results of this research show a wide variability in the morphological characteristics of bean seeds. This diversity is usually common in traditional varieties that have faced less selection processes than the commercially developed ones [20]. These results are consistent with the findings of other studies [18,71]. Freitas et al. [18] assessed 50 accessions of common bean, representing the diversity cultivated on the Island of Madeira, an area with a wide range of ecological conditions. The common bean materials evaluated in that study showed a broad intraspecific variability. It is important to mention that those materials had, presumably, their origin in the Andean gene pool, based on their results of morphological traits. On the other hand, Loko et al. [71] evaluated 57 common bean accessions from the Benin Republic based on seed morphological traits. As expected, they found a wide range of variability among landraces as well.

Landraces with larger populations were Canario (present in 31 out of 51 chakras) and Poroto Grande (28 out of 51 chakras). In contrast, Poroto Campeon and Sucu Tomate Pintado were found to have limited populations, being restricted to a single side each (

Table 3. Bean landrace names given by smallholder farmers and their respective translations from Kichwa and name meaning.

Landrace	Translation, Name Meaning	Photo 1	AF 2	Landrace	Translation, Name Meaning	Photo 1	AF 2
(Farmers’ Name)				(Farmers’ Name)
Canario 3	“Canary” (seed color)	7	31	Puca Poroto	“Red bean” (seed color)	20	6
Poroto Grande	“Big bean” (seed size)	17	28	Yura Pintado	“White spotted” (seed color)	38	6
Popayan Morado	“Purple bean” (seed color)	43	14	Bolon Pintado	“Round mottled” (seed shape, color)	3	5
Josico	Unknown meaning	11	13	Bolon Rojo	“Red round” (seed color, shape)	4	5
Yana Poroto	“Black bean” (seed color)	34	13	Paco Bolon Rayado	“Buff, round, striped” (seed color, shape, pattern)	42	5
Puca Pintado	“Red mottled” (seed color)	18	12	Poroto Conejo	“Rabbit bean” (seed appearance)	16	5
Sara Poroto	“Corn” (crop association)	24	12	Racu Pintado Poroto	“Thick speckled bean” (seed size, pattern)	22	5
Yura Allpa Poroto	“Ground white bean” (habit, seed color)	37	12	Rayado Poroto	“Stripped bean” (seed pattern)	23	5
Allpa Poroto	“Ground bean” (habit)	1	11	Yana Pintado	“Black mottled” (seed color, pattern)	33	5
Killu Canario	“Yellow canary” (seed color)	13	11	Sucu Allpa Poroto	“Ashen ground bean” (seed color, habit)	25	4
Cargabello	Juxtaposed word meaning: high yield	8	10	Sucu Pintado	“Ashen spotted” (seed color, pattern)	26	4
Frejol Tomate	“Orange bean” (seed color)	10	10	Sucu Rayado	“Ashen striped” (seed color, pattern)	29	4
Popayan Morado Rayado	“Purple stripped bean” (seed color, pattern)	44	10	Yura Popayan	“White bean” (seed color)	47	4
Sucu Poroto	“Ashen bean” (seed color)	27	10	Caca de Conejo	“Rabbit excrement” (seed appearance)	5	3
Yana Vaca Poroto	“Black spottled bean” (seed color, pattern)	36	10	Frejol Duro	“Hard bean” (seed consistency)	9	3
Popayan Pintado	“Speckled bean” (seed pattern)	45	9	Tomate Vaca Poroto	“Orange cow bean” (seed color, pattern)	32	3
Sucu Poroto Pintado	“Ashen spotted bean” (seed color, pattern)	22	9	Yura Pintado Poroto	“White spotted bean” (seed color, pattern)	39	3
Café Pintado	“Brown speckled” (seed color, pattern)	6	8	Bolon Morado	“Purple round” (seed color, shape)	2	2
Matambre Negro	“Black kill hunger” (seed color, juxtaposed word meaning: kill hunger)	14	8	Killu Allpa Poroto	“Yellow ground bean” (seed color, habit)	12	2
Yana Sucu Poroto	“Orange black bean” (seed color)	35	8	Puca Vaca Pintado	“Red cow spotted” (seed color, pattern)	21	2
Paco Bolon	“Light yellow brown, round” (seed color, shape)	41	7	Sucu Vaca Poroto	“Ashen cow bean” (seed color, pattern)	31	2
Poroto Pintado	“Mottled bean” (seed pattern)	19	7	Poroto Campeon	“Champion bean” (yield)	15	1
Yura Poroto	“White bean” (seed color)	40	7	Sucu Tomate Pintado	Red orange speckled (seed color, pattern)	30	1
Paco Popayan Pintado	“White speckled bean” (seed color, pattern)	46	6

¹ See Figure 3; ² AF = absolute frequency; ³ Canario is not considered a landrace; instead, it is an improved variety.

). Farmers expressed their preference for cultivating Canario, especially motivated by commercial purposes and the consumer demand it generates. It is important to mention that Canario is not a landrace; rather, it is an improved variety.

3.3. Morphological Characterization

Concerning the primary color of the bean seed, we found a wide variability among landraces, which varied from light to dark colors. The most frequently observed were pale cream to buff (23%), white (16%), and orange (13% of the 361 samples collected), with smaller percentages found in black and gray colors. As mentioned by Abera et al. [1], seed color is the main morphological characteristic used to identify bean landraces. This attribute not only allows a better communication between researchers and farmers but also promotes improved mutual understanding.

Regarding the secondary seed color, 41% of the samples showed no color, 29% had black as the secondary color of the seed, and 7% of the samples showed brown and red colors. Brown was the most common color around the hilum (47% of the samples). A total of 41% of the samples had no seed coat pattern. Regarding the seed shape, the most frequent one was kidney-shaped (51%).

The mean, standard deviation, coefficient of variation, and minimum and maximum values of three quantitative variables can be found in the Supplementary Materials. The results reveal that seed thickness exhibits greater variability (CV = 15.21) compared to the other morphological traits, as evidenced by its higher coefficient of variation.

In the Supplementary Materials are presented summary statistics for the five quantitative bean traits. Our results show that the mean values for the attribute seed length (L) varied from 10.3 mm (Yura Allpa Poroto) to 19.5 mm (Paco Bolon). The seed width (W) values ranged from 4.2 mm (Sucu Allpa Poroto) to 13.6 mm (Paco Bolon), and seed thickness (T) values varied between 3.9 mm (Matambre Negro) and 9.5 mm (Canario). The lowest seed length/width ratio (L/W) was 1.1 mm (Canario), while the highest ratio registered was 3.3 mm (Sucu Allpa Poroto). Regarding the seed width/thickness ratio (W/T), the values differed from 0.8 mm (Sucu Allpa Poroto) to 2.2 mm (Paco Bolon).

The coefficient of variation for the variable seed length (L) was higher for the landrace Sucu Pintado. This landrace presented the greatest variability in the characteristic seed thickness (T) as well. Regarding the variable seed width (W), the Sucu Allpa Poroto landrace had the highest heterogeneity, which is considered a high intra-varietal variability.

The results of the cluster analysis revealed that the 47 landraces could be clustered into three groups (Figure 3). The cophenetic correlation coefficient was 0.6.

The first group (G1) represented by 22 landraces showed seed average dimensions as follows: L = 13.84 mm, W = 10.04 mm and, T = 7.41 mm, which were the highest within the three groups; the aspect ratio values, L/W and W/T, were 1.38 and 1.37, respectively; the accessions in the group displayed mainly white seed primary color, no secondary color, brown color around the hilum, no pattern, and an oval shape. The seeds of G1 were the largest and roundest of all the groups.

The second group (G2) contained six landraces with the following seed average dimensions: L = 13.51 mm and W = 8.27 mm, which were the lowest among the groups, and T = 6.20 mm, L/W = 1.61, and W/T = 1.34; the seeds in this group exhibited a primary color ranging from pale cream to buff, generally without a secondary color or pattern, brown color around the hilum was predominant, and most of the seeds were kidney-shaped.

The accessions within the third group (G3) included 19 landraces, characterized by the following seed dimensions: L = 13.69 mm, W = 8.29 mm, T = 6.06 mm—this last one was the lowest of all three groups—L/W = 1.67, and W/T = 1.39; the main seed primary color was pale cream to buff, black was the principal secondary color and color around the hilum, and most of the accessions showed a striped pattern and kidney shape.

Principal component analysis showed that 94% of the variation was explained by PC1 and PC2 (Figure 4 and

Table 4. Eigenvalues and the contribution of each principal component based on the seed length (L), seed width (W), seed thickness (T), L/W ratio, and W/T ratio.

Component	Eigenvalue	Contribution Rate (%)	Accumulative Contribution Rate (%)
PC1	2.53	51%	51%
PC2	2.17	43%	94%
PC3	0.29	6%	100%

). The first principal component (PC1) explained the highest percentage of variance (51%), and it was correlated with seed thickness (T) and the ratio L/W. The second principal component (PC2) justified 43% of the variation and was highly correlated with the seed length (L) and the ratio W/T. Most of the landraces of G1 were located on the right side of the plot (red circles). The P. coccineus landraces were concentrated in the upper right quadrant (blue circles), exhibiting a distinct boundary from the other groups. Landraces from G2 and G3 were placed on the left side of the plot (green circles). There were also outliers in the group, represented by landraces Poroto Campeon, Sucu Allpa Poroto, Yana Pintado, and Rayado Poroto.

Significant differences (p < 0.05) were detected on the traits of seed width (W), seed thickness (T), and the L/W ratio (

Table 5. ANOVA for the five bean seed quantitative traits.

Trait	p-Value
Seed Lenght (L)	0.3473
Seed Width (W)	<0.0001
Seed Thickness (T)	<0.0001
L/W	<0.0001
W/T	0.3439

). On the other hand, we did not find statistical differences on the seed length (L) trait, nor the W/T ratio.

3.4. Geographic Distribution

As shown in Figure 5, we found bean landraces distributed in the three administrative zones. The number of landraces in each chakra ranged from 1 to 6 (26%), 7 to 12 (24%), and 13 to 18 (8%).

As seen in Figure 6, we found a positive spatial autocorrelation (Moran’s I: 0.24; z-score: 2.20; p-value: 0.03). Since the index value is positive, we assume that the number of landraces is associated with close locations [85]. Given the z-score of 2.20, there is a less than 5% likelihood that this clustered pattern could be the result of random chance, which confirms the presence of a spatial structure.

3.5. Diversity Analysis

The findings show that the values of the Shannon index (H′) vary within a range of 0.45 to 1.55 (

Table 6. Shannon–Wiener Diversity–Equity (H′) index and Shannon J’ homogeneity index for P. vulgaris and P. coccineus landraces at the community level.

Index	Arrayanes	Chilcapamba	Cumbas Conde	El Cercado	Italquí	Morlán	Morochos	San Antonio Punge	San Pedro	Topo Grande
Shannon H′ Log Base 10	1.322	0.987	1.26	1.455	0.941	1.315	1.549	0.452	1.246	1.002
Shannon J’	0.971	0.987	0.953	0.943	0.941	0.965	0.96	0.946	0.958	0.962

). The Morochos community exhibits the highest P. vulgaris and P. coccineus diversity among the communities.

In terms of homogeneity (Shannon J’), the Chilcapamba community exhibits the highest value (0.99), suggesting that the P. vulgaris and P. coccineus landraces in this community are more similar to each other. This implies a reduced diversity within the community.

4. Discussion

We found a wide variability of beans (Figure 2) distributed in the study area (Figure 5) (46 landraces and a commercial variety) regularly linked to traditional reasons. This diversity was predominantly kept by indigenous female farmers with limited access to formal education. Their agriculture depends on weather conditions, as most of them do not have access to irrigation water. Considering that bean diversity in the study area is conserved by a disadvantaged population, this study highlights the need for public policies that improve their living conditions and, consequently, indirectly favor agrobiodiversity conservation.

The Canario variety was the most abundant within the studied chakras (Table 3). It was geographically distributed in the three administrative zones. The relatively high abundance of this variety might be explained by its commercial purpose. Most of the interviewees stated that they chose to cultivate this particular landrace due to its high demand among consumers. On the other hand, the second-most abundant landrace was Poroto Grande. Similar to Canario, this landrace was distributed throughout the study area. Coincidentally, both share a similar color tone of their seeds. Canario was classified as light yellow, while Poroto Grande showed a medium yellow brown color. This suggests that yellow tones represent a desirable trait among farmers and consumers. Other authors have reported a preference for yellow tone common bean as well [86,87,88]. The interest in yellow beans can be attributed to their better digestibility characteristics, as well as the less intestinal distress they cause [87,89]. As a complement, it is worth mentioning that, although probably not linked to consumers’ preference, research has revealed that yellow bean varieties have a higher protein content compared to red seeds [90]. Being a commercial variety, numerous farmers cultivate it for trade due to their preferences and its superior yield. Consequently, this could represent a high risk of genetic erosion, as landraces and other low-yielding cultivars might eventually be displaced. This finding leads to exploring strategies for diversifying the production and promoting the cultivation of traditional varieties.

Consistent with previous studies [91,92], our findings also highlight the significance of seed shape and color combination in influencing the preferences of both farmers and consumers. The significant diversity registered in seed color suggests that farmers’ preferences probably influenced the distribution and conservation of these landraces based on seed colors.

Our findings suggest that Poroto Campeon and Sucu Tomate Pintado landraces are the ones that are, in fact, facing a substantial risk of disappearing. This result also implies that farmers’ perception may not always be entirely accurate, and further information should be considered to contrast the information given by farmers.

The cophenetic correlation coefficient of 0.6 in the cluster analysis indicates reasonably well-defined clusters based on the morphological seed traits within the three groups (Figure 3). The landraces in G1 were the most abundant (22), and their seeds showed the largest dimensions of all the groups, exhibiting an oval shape and mainly white seed primary color with no pattern. The seeds in G2 represented the smallest group (six landraces) and showed the shortest length and width dimensions, pale cream to buff primary color generally without a pattern, and most of the seeds were kidney-shaped. The seeds in G3 included 19 landraces and presented the lowest thickness of all the groups, the main seed primary color was pale cream to buff, and most of the accessions showed a striped pattern (with black as the principal secondary color) and kidney shape. These results could lead to selecting landraces for future breeding programs to obtain desirable characteristics based on the consumer preferences.

The principal component analysis revealed that most of the landraces of G1 were located on the right side of the plot (Figure 4). The P. coccineus landraces were concentrated on the upper right side, clearly separated from the other groups. In contrast, the landraces from G2 and G3 were located on the left side of the plot. The first principal component (Eigenvalue = 2.53) explained the highest percentage of variance (51%) and was correlated with the seed thickness (T) and the L/W ratio (Table 5). This result should be considered for further study due to its contribution to the total diversity.

The positive Moran’s index indicates that chakras with higher diversity tend to be distributed closer to each other, as do those with lower diversity (Figure 6). This clustering tendency suggests that there are factors that might be influencing this spatial distribution [93]. In this study, the factors potentially influencing this distribution may likely involve seed exchanges among neighbors.

The highest Shannon diversity index (H′) in this study was found in the Morochos community (H′ = 1.55) (Table 6). It was higher compared to the reports of [69] on common bean (H′ = 0.74). The high diversity found in this community makes it a potential site for both sourcing genetic resources and implementing conservation programs.

This diversity could be attributed to the traditional practices of indigenous communities in the study area, where landraces are usually kept in the chakras and seeds are frequently passed down from one generation to the next.

5. Conclusions

This study describes the morphological seed characteristics of 361 georeferenced samples of bean landraces cultivated in three administrative zones: Imantag, El Sagrario, and Quiroga in the Andes Highlands of Cotacachi Canton in Ecuador. Based on the seed morphological descriptors, as well as the information provided by farmers, we identified 46 bean landraces and 1 commercial variety. The landraces with a wider distribution and population throughout the study area were Canario and Poroto Grande, probably due to the preference by farmers and consumers for yellow common bean. Contrary to farmers’ perception, our findings revealed that landraces Poroto Campeon and Sucu Tomate Pintado had limited distribution and scarce populations in the study area. This observation raises awareness about the potential vulnerability of these landraces, representing a possible risk of their disappearance. The high diversity found in the Morochos community (H′ = 1.55) makes it a promising geographic area for sourcing germplasm for further study. The diversity in the studied common and runner bean landraces, as well as their distribution, was confirmed by morphological traits, diversity indices, spatial autocorrelation, and cluster and principal component analyses, and the results represent a contribution to further research focused on the conservation of bean diversity in the Andean area of Cotacachi.