Traditional knowledge of the diversity of chilli Amashito

In total, 98 populations of chilli Amashito were collected in the plain of the state of Tabasco and northern Chiapas, these populations were classified by and the collectors into nine morphological variants (morphotypes) of wild chilli Amashito (Capsicum annuum var. Glabriusculum) based on mainly immature fruit color, shape and size characteristics: Amashito verde, Amashito ojito de cangrejo, Amashito blanco, garbancillo blanco, garbancillo verde, Amashito grande or gigante, Amashito bolita, Amashito muela and colmillo de lagarto ( (Table 1).

That is, the gatherers name the variants of the chilli Amashito based on their characteristics of the fruit. In general, the fruits of the Amashito verde, garbanzo, ojito de cangrejo and bolita morphotypes are mainly characterized by having almost round shapes. In this sense, according to ^{Hernández-Verdugo et al. (1998)}; ^{Mongkolporn and Taylor (2011)}; ^{Bosland and Votava (2012)}; ^{Hayano-Kanashiro et al. (2016)}; ^{Velázquez-Ventura et al. (2018)}, mention that the almost round shape of the fruit is one of the main features presented in the wild chili (Capsicum annuum var. Glabriusculum).

In previous studies carried out in the state of Tabasco, only 3 to 4 morphotypes have been reported: Amashito verde, ojitos de cangrejo, garbanzo and muela (Castañón-Nájera et al., 2008; ^{Prado-Urbina et al., 2015}; Velázquez -Ventura et al., 2018) while, in Chiapas, ^{Bran et al. (2012)} mention wild chili as Timpinchile without specifying the variation between them. This indicates that the phenotypic variation of this type of chili in the study regions is possibly not well understood. In ethnobotany, the morphotype is the first level of diversity, but there is also intra-morphotype diversity that can be exploited for conservation and use purposes (^{Latournerie et al., 2002}).

It was observed that the diversity of chilli Amashito by region (municipality) was greater in the municipalities of Huimanguillo, Cunduacan and Comalcalco in the state of Tabasco with 7, 6 and 5 morphotypes, respectively (Table 2). Which is also related to the regions where the greatest number of collections were obtained (between 8 and 12). While a smaller number of variants was found in the municipalities of Juárez and Salto de Agua that correspond to northern Chiapas, with 2 morphotypes of chilli Amashito each.

Regarding the abundance by morphotype type of chili Amashito collected, the largest number of populations collected correspond to the Amashito verde (36.7%), followed by ojito de cangrejo (12.3%), Amashito blanco and garbancillo blanco with 11.2% each. On the other hand, the colmillo de lagarto morphotype was only found in two localities (common Malpasito of the municipality of Tecpatan, Chiapas and in the Libertad municipality of Huimanguillo, Tabasco) that are located very close to each other, so they could be lost due to the low frequencies in which they are found in a region (Table 2).

In this regard, ^{Ramírez-Meraz et al. (2015)} point out that it is evident that there is a risk of the potential disappearance of this wild chili, because the natural reproduction of this genetic resource is difficult since the area of native vegetation of the regions where the chili grows is reduced every year and with it the natural habitat of many birds that participate in its dispersion is destroyed.

The results of the interviews carried out with 39 collectors of the study sites, 100% of them identify the Amashito verde (Table 3), while 35.9% know the ojito de cangrejo and the Amashito blanco and 25.6% the garbancillo blanco. In other words, these four morphotypes of chilli Amashito are the best known in the study regions. The lesser-known morphotypes of the inhabitants are the Amashito bolita (10.3%) and the colmillo de lagarto (2%), a result that coincides with the few samples collected from these morphotypes in the study area (Table 2).

On the other hand, the most common morphotypes identified by the collectors in this study, mostly coincide with those reported by Castañón-Nájera et al. (2008); ^{Velázquez-Ventura et al. (2018)}; ^{Gálvez et al. (2018)}, which were Amashito verde, ojito de cangrejo, garbanzo and muela. This indicates that these morphotypes are the best known and present in the studied areas.

On the other hand, the collectors identify nine ecosystems (habitat) where the chilli Amashito grows (Table 3): cocoa farm, acahual, banana trees, cornfields, paddocks, roadside, fields, fruit orchards and citrus. Being more frequent to find the chilli Amashito in the cacao farms (82.1%), acahuales (74%), plantains and cornfields (53.8% each) and paddocks (46.2%). The ecosystems where the least presence of chilli Amashito, less than 30%, are located are fields, and fruit and citrus orchards. Presenting the Amashito verde in most ecosystems. Furthermore, an apparent relationship between ecosystems and some particular morphotype was not observed.

Phenotypic diversity analysis

Table 4 shows the results of the principal component analysis (PCA) with the 98 populations of chilli Amashito with the 13 selected morphological variables. The first three main components explain 50.2% of the observed morphological variability. The PC1 contributed to explain 22.8% of the total variation and an eigenvalue of 3, mainly associated with the variables leaf shape (FH), leaf length and width ratio (LH/AH) in a positive way and in a negative way with leaf margin, ratio of plant height and stem diameter (AP/DT).

In other words, this component was defined to a greater degree by variables of the plant structure. This coincides with studies of morphological diversity of wild chilies, where it is reported that the first main component was explained by leaf and plant descriptors (^{Castañón-Nájera et al., 2010}; ^{Moreno-Pérez et al., 2011}; ^{de la Cruz-Lázaro et al., 2017}).

In this sense, ^{Villota-Cerón et al. (2012)} report that the first three main components contributed to explain 61.9% of the variation observed in different types of chilies (Capsicum spp.), While ^{Gálvez et al. (2018)} in in situ morphological diversity studies with 9 descriptors of wild and semi-wild populations (C. annuum var. glabriuculum and C. frutescens) found that the first three main components explained 58.3% of the variation.

According to the first two main components (PC1 and PC2), the germplasm was distributed in the four quadrants and four large groups were formed, in addition to two collections that were separated from the rest, forming an individual group each (Figure 2). Group 1 (G1) was formed with 22 populations of chilli Amashito distributed in quadrants I and IV. These were collected in the most remote communities in the mountainous area of Chiapas (560 to 800 masl), in the municipalities of Ostuacan, Malpaso and Juárez. The collections are characterized by presenting plants with greater height (between 1.4 to 2.4 m), large leaves with an average length of 4.4 cm and 6.2 cm in width, elongated fruits of greater size (from 1.9 to 2.6 cm in length and diameter of 0.7 to 2.5 cm).

Figure 2 Distribution of 98 populations of chilli Amashito (C. annuum var. Glabriusculum) from Tabasco and northern Chiapas, Mexico according to the two main components. 

In general, it was observed that the variation explained in the chilli Amashito due to its morphological characteristics of the plant and fruit of the northern regions of Chiapas and the plain of the state of Tabasco are distributed in distant groups, indicating diversity based on the region where they are distributed, given that each region is physically separated and distant with different physiographic characteristics and conditions in altitude, soil and vegetation, so the Tabasco region is on a plain and that of northern Chiapas in the hills and mountains.

In this regard, ^{Hernández-Verdugo et al. (2012)} found a positive correlation between morphological variation and climatic factors (temperature and water) at the sites of origin of the studied populations of C. annuum var. Glabriusculum. Furthermore, they point out that the populations of wild C. annuum in northwestern Mexico have high variation in their morphological characteristics measured within and between them, indicating that this species is a valuable genetic resource that must be studied to improve its use and conservation.

Group 3 (G3) consisted of 19 populations that were distributed in the first and second quadrant (Figure 2), formed by germplasm that comes from both studied regions and that are located in neighboring communities (communities of the municipalities of Tacotalpa, Huimanguillo and Cardenas de Tabasco and by Ostuacan, Malpaso and Pichucalco of northern Chiapas), so the flow of commercial activities among the inhabitants of these communities, as well as birds, among others, favors the flow of biological material, which it leads to these populations showing a certain relationship. Group 4 (G4), located in the third and fourth quadrant, was formed by 21 populations of communities from the municipalities of Cunduacan, Jalpa and Nacajuca, in the state of Tabasco, with altitudes less than 10 m.

Population 56 formed a single group (G5), this population was collected in the La Granja Belen ranch in the municipality of Macuspana, Tabasco, it has distinctive characteristics of average plant heights of 1.4 m, leaves with averages of 5 cm in length and 4 cm wide, medium fruit size (1.5 cm long and 0.8 cm in diameter) and fruit color at maturity light red. On the other hand, in the Topotzingo community, Nacajuca municipality, Tabasco, the 49th population that formed the G6 group was collected.

Based on the criteria of group formation (^{Núñez-Colín and Escobedo-López, 2011}), the analysis of conglomerates at a Euclidean distance of 4, was complemented by what was observed in the analysis of principal components, due to the similarity in the way each analysis grouped the evaluated populations. Figure 3 shows that a large group (G1) was formed that included the Chiapas and Tabasco germplasm (differentiated into four well-integrated subgroups: S1G1, S2G1, S3G1 and S4G1) and populations 49 and 56 that each individual groups formed.

Figure 3 Dendrogram of 98 populations of chilli Amashito (C. annuum var. Glabriusculum) collected in the state of Tabasco and northern Chiapas. 

Subgroup 1 (SG1) was made up of eight populations from the Tabasco region and they are distinguished by having almost round small fruits that grow mainly on cocoa and milpa farms. SG2 made up of 29 populations that are distributed in neighboring communities between the Chiapas and Tabasco regions were found in ecosystems of cacao and acahual estates. These populations are distinguished by presenting small and mainly round fruits. In S3G1, 20 populations with distinctive characteristics of larger and elongated fruits were grouped, as well as plants of greater height, which were found in the mountainous area of northern Chiapas (560 to 800 masl).

S4G1 was made up of 18 populations from the lower Tabasco zone (less than 10 masl). Furthermore, within group 1, there were 6 independent populations (32, 13, 46, 79, 10 and 1) that did not form a defined subgroup, most of them from the state of Tabasco. Populations 56 and 49 that were collected in Tabasco defined groups 2 and 3, respectively.

In general, the classification of diversity at the level of morphotypes handled by collectors in the study sites of the chilli Amashito, did not coincide with the way in which the variation in the phenotypic diversity study was grouped. This is explained by the fact that the gatherers name the variants based on distinctive characteristics of personal appreciation, mainly of the fruit, such as the shape of the anthocyanin stains of the fruit, for example: crab eye, the shape of the fruit with a pointed finish, for example: colmillo de lagarto, shape of the base of the fruit, for example: muela, shape of the fruit almost round, for example, Amashito bolita, among others; which does not correspond to the descriptors for Capsicum (^{IPGRI, AVRDC and CATIE, 1995}).