Introduction

Biodiversity results from multiple historical and abiotic factors, ecological interactions, and evolutionary processes (^{Futuyma, 2013}). Some living organisms, such as plants, have been of interest to humans; one result of this interaction is that humans have constantly used and selected them. These actions have driven diversity in cultivated or exploited plants through manipulation and artificial selection as part of the evolutionary process of domestication (^{Pickersgill, 2007}; ^{Zeder et al., 2006}). One consequence of domestication is morphological diversity, the raw material on which peasants rely for selecting plant varieties because morphological features are conspicuous. When artificial selection fixes some characters, making morphological differences noticeable, peasants usually assign different names to denote groups of plants with characteristics (^{Berlin, 1992}; ^{Pickersgill, 2007}; ^{Zeder et al., 2006}). These local names may correspond to landraces. Theoretically, landraces are cultivated plants that differ from their wild relatives by having adaptations that allow them to grow outside of their natural habitat; landraces can also show morphological changes known as “domestication syndrome”, which reflects artificial selection (^{Casañas et al., 2017}; ^{Pickersgill, 2007}).

Mexico is a diversity center of the genus Agave L. since it is home to 79% of the species, most being endemic to the country (^{García-Mendoza, 2011}). In addition, the diversity of the genus Agave is also due to the use, manipulation, and human selection of agave plants (^{Eguiarte et al., 2021}; ^{Gentry, 1982}). Humans have been using them for at least 10,000 years; virtually the whole plant is harnessed, and more than 92 uses have been reported (^{Colunga-GarcíaMarín et al., 2017}; ^{Trejo et al., 2022}). The uses of agave include the elaboration of distilled spirits such as tequila or fermented beverages such as “pulque”. The latter is a traditional beverage obtained from the fermentation of Agave sap (“aguamiel”) that has been prepared since pre-Hispanic times that has recently regained relevance. The state of Tlaxcala ranks second in Mexico in “pulque” agave and “aguamiel” production, after the state of Hidalgo (^{SIAP, 2018}).

Studies on the morphological diversity of agave plants have been based on measuring all possible morphological -mainly vegetative- characteristics. Reproductive traits are hard to observe because the floral scape is removed by peasants for the consumption of flowers and the floral scape but mainly to stimulate the accumulation of sugars in the stem in order to produce sap (^{Figueredo-Urbina, Álvarez-Ríos, & Zárraga-Cortés, 2021}). Little research has been conducted on how peasants identify their landraces (^{Colunga-García Marín & May-Pat, 1993}; ^{Ramírez-Manzano et al., 2020}). Traditional knowledge can be essential in understanding agave diversity and its management and selection processes.

The present work explored the diversity of agave plants in the Nahuatl locality of San Pedro Tlalcuapan, Tlaxcala, focusing on the following objectives: i) identify species, subspecies, and landraces; ii) characterize the morphological attributes used by peasants to identify landraces; iii) integrate conventional statistical and traditional-knowledge approaches to supplement the description of agave diversity in the study area.

Materials and methods

The study was conducted in San Pedro Tlalcuapan (SPT), a Nahuatl indigenous community located northwest of La Malinche National Park (PNLM, for its acronym in Spanish) (19°16’50.02” N, 98°09’06.30” W), in the municipality of Chiautempan, state of Tlaxcala, Mexico (Fig. 1). SPT is located at an elevation of 2,411 m, with annual mean precipitation and temperature values of 901 mm and 15.7 °C that characterize a subhumid temperate climate C(w₁) (^{García, 1988}). The vegetation is made up of coniferous forest (Abies and Pinus) and mixed (Pinus-Quercus and Quercus-Pinus) temperate forests under different levels of conservation, as well as induced pastures and seasonal crops alternating with rows of agave plants, locally known as “metepantles” (^{INEGI, 2010}).

Figure 1 Geographic representation of the Nahuatl indigenous community of San Pedro Tlalcuapan, municipality of Chiautempan, Tlaxcala, Mexico. 

The name ‘Tlalcuapan’ derives from the Nahuatl (“tlali” land, “cuapan” - header), which means “at the head of the earth” or “on the slopes of the Malintzi”. San Pedro Tlalcuapan has a population of 3,613 inhabitants living in 837 households: of these, 1,313 (36.34%) inhabitants are indigenous persons with a medium marginalization level (^{Secretaría de Bienestar, 2010}), 45% are between 20 and 59 years old and 10% are over 60 years old (^{INEGI 2010}). Previously, the local economy was based almost entirely on forest natural resources, including mushrooms, plants, and animals. This relationship led to the acquisition of knowledge, practices, and other cultural elements. One of the occupations is “tlachiquero” (subfix-ero; “tlachique” the one who scrapes), the person who knows how to extract sap from agave plants; “pulque” has been produced since at least 500 years ago. Currently, there is a decreasing trend in the abundance of forest natural resources, along with disinterest in the ancestral culture and changes in traditional knowledge, likely associated, partially with the degradation of the local ecosystems (^{Bello-Pérez, 1994}; ^{Bello & Pérez, 2019}).

To address this situation, community members and local authorities have focused on revaluing, reaffirming, and supporting their biocultural heritage conservation and fostering sustainable development. Tlalcuapan is a community governed under a ‘uses-and-customs’ regime, i.e., it is ruled by its own laws and regulations. The highest authority is the community assembly, which has maintained autonomous government and resource management (^{Bello-Pérez, 1994}; ^{Bello & Pérez, 2019}).

Identification of landraces through traditional knowledge

Interviews were administered to agave producers, and field trips were conducted with them from July to December 2016 in SPT, following the guidelines of the Code of Ethics for Ethnobiological Research in Latin America (^{Argueta et al., 2016}). The support of some members and community representatives allowed us to approach the local authorities and persons familiar with the management of agave plants. These persons were informed about the objectives and activities to be performed, requesting their support and participation in the research.

A total of 10 semi-structured interviews were administered, and 3 field trips were conducted. Five interviews were applied to “pulque” producers (“tlachi- queros”) and 5 to non -“tlachiqueros”. “Pulque” producers were selected through a non-probabilistic sampling known as the snowball method (^{Bernard, 2006}), which consists of locating 1 or 2 persons who lead to others, and so on. All the local experts were interviewed since few persons currently work as “tlachiqueros”. Three visits were conducted based on the availability of these experts; each guided us through their land and showed us the different agave landraces.

During the interviews, 3 variables were included following ^{Turner (1988)}: frequency of mentions (number of times a group of people mentions a plant), lexical marking (naming a plant taxon; this name does not vary across groups of people when the plant is of great cultural importance), and perception (including factors such as plant size or attributes that confer notoriety to agave landraces in a locality, including leaves, flowers, and fragrance, among the characters of importance for identification).

The qualitative data collected were sorted into 2 categories: common name mentioned and plant morphology. This information was analyzed using descriptive statistics, percentage of each Agave landrace recorded, and morphological characters used to identify landraces. Data matrices were built for the frequency of mentions using a binary encoding (1 = mentioned; 0 = not mentioned). In the case of the morphology category, descriptive statistics were applied only to the 3 landraces with the highest frequency of mentions, which coincided with the most abundant ones (“Manso”, “Prieto”, and “Amarillo”). The perception analysis considered all terms used by peasants to describe the morphological traits of landraces; the result was simple adjectives not based on any parameter established in official metrics or tables. Three categories were assigned for plant size and rosette length, 3 categories were assigned: large = 3, medium = 2, small = 1; 2 for rosette width: thick = 2, thin = 1; 2 for spine stiffness (terminal spine): flexible = 2, stiff = 1; and 3 for plant color: dark green = 3, yellow = 2, ash green = 1.

The common names (landraces) used by local experts to identify the agave varieties were recorded; subsequently, the scientific name of each variety was determined using taxonomic keys (^{Gentry, 1982}; ^{García-Mendoza, 2011}). In addition, leaves were collected to prepare herbarium specimens deposited at the National Herbarium of Mexico, MEXU, UNAM, and the Herbarium of Universidad Autónoma de Tlaxcala TLXM, UATx.

Morphological diversity between landraces

To explore the morphological diversity of the agave landraces used by the community of study, adult plants at least 8 years old were selected from the most abundant landraces located in “metepantles” (rows of agave plants alternating with rows of other cultivated species) grown in corn, broad bean, and common bean plots. Since less than 3 individuals were found for most landraces, we measured 25 individuals of “Manso” and “Prieto” and 11 of “Amarillo”. Thirty-two morphological characters (https://zenodo.org/record/3976297#.Xy35gChKjIU) selected according to ^{Gentry (1982)}, ^{Colunga-GarcíaMarín and May-Pat (1996)}, and ^{Vargas-Ponce et al. (2007)} were recorded for each plant; 25 were quantitative and 7, qualitative. Measurements were made with a vernier for small attributes (mm), a tape measure for larger attributes (cm, m), and a micrometer (to measure the cuticle). Color was recorded using the Munsell color chart (1991), and samples of whole leaves were collected to obtain morphometric data. Only vegetative characters were considered due to the absence of plants with reproductive structures.

Qualitative multistate characters were transformed by assigning an ordinal function; a basic data matrix consisting of 61 rows (individuals) and 32 columns (morphological characters) was built using the collected information. Morphological character data were standardized; then, a correlation analysis was run to identify highly correlated variables and exclude them from the matrix. After the correspondence analysis, only 11 characters remained. A similarity matrix was generated using the Euclidean distance and a dendrogram of morphological distances was built with the farthest neighbor method to assess the similarity between specimens of the 3 local landraces. Subsequently, a principal component analysis (PCA) was carried out to arrange the specimens according to their morphological characteristics and identify redundant variables and those with the highest variation that dominated the arrangement.

Finally, a multivariate analysis of variance was performed using the values of the 11 morphological variables, followed by a Tukey’s test to identify those variables with statistically significant differences between landraces. All statistical analyses were performed in the R platform (^{R Core Team, 2016}).

Results

In the locality of SPT, 2 species and 8 landraces of agave are recognized. Among these, Agave americana L. landraces “Palma” and “Cenizo” are identified by interviewees as introduced, whereas Agave salmiana Otto ex Salm-Dyck subsp. salmiana landraces “Amarillo”, “Chino”, “Mahuahuac”, “Manso”, “Matecon”, and “Prieto” are considered native to the locality (Fig. 2).

Figure 2 “Local varieties” (landraces) of Agave in San Pedro Tlalcuapan. Agave salmiana subsp. salmiana: A) “Manso”, B) “Prieto”, C) “Amarillo”, D) “Mahuahuac”, E) “Chino”, G) “Matecon”. Agave americana: F) “Palma” y H) “Cenizo”. The characters marked with (*) are used by peasants to recognize landraces; the characters marked with (**) are used by peasants and supported with average morphological data from statistical analyses. Data with no asterisks correspond to taxonomic observations. 

Recognition of landraces through traditional knowledge

The most frequently mentioned landraces are those with the highest abundance in the area. “Manso” ranks first in frequency of mentions, being the most abundant landrace in the area. “Amarillo” ranks second despite being less abundant than “Prieto”, which is in third place (Fig. 3).

Figure 3 Frequency of mentions of landraces or local varieties in San Pedro Tlalcuapan. 

On the lexical marking, it was found that some landraces are named in Nahuatl because San Pedro Tlalcuapan is a community that still preserves indigenous traditions. This is the case of the following landraces: “Manso” - Teometl (teotl = deity, metl = agave), “Amarillo” - Cosmetl (costic = yellow, metl = agave), “Prieto” - Xilumetl (xilutl = tender, metl = maguey), “Cenizo” - Nexmetl (nextic = ash-colored, metl = maguey), and “Chino” - metl chinte (metl = agave, chinte = Chinese).

Finally, regarding perception, we observed 7 qualitative characters used by experts to identify traditional varieties (landraces). According to the analyses, 5 qualitative characters have the highest number of mentions to identify the 3 most abundant landraces of Agave salmiana subsp. salmiana (Fig. 2). However, specific characters are also used for identifying each landrace. “Manso” is determined by peasants as a large agave, more prominent than “Prieto” and “Amarillo”. It has long, thick leaves and stiff spines, and the color of the plant is ash-green. “Amarillo” is identified as being smaller than “Manso” but larger than “Prieto”, with medium-sized thin leaves. Its distinctive traits are its flexible spines and yellowish-green leaves. “Prieto” is described as a small-sized agave with thin, small, dark-green leaves and stiff spines.

Today, all the landraces reported in SPT are used mainly for the manufacture of “pulque”, but in the past these plants were also used for the extraction of a range of materials, including fibers, “mixiote”, and soap, among others; in addition, agave played a role in the protection of cropland and the delimitation of parcels. These uses and the discontinuation of some of them may explain the presence of the distinctive characteristics of the different varieties and the decline of some populations. For example, the “Prieto” landrace is mainly used to delimit parcels, in addition to serving to protecting the soil and storing moisture; this smaller landrace optimizes space for the free movement of animals in the field, which would explain the selection of this characteristic.

Morphological diversity across landraces

The correlation analysis allowed excluding those variables that provided little information in the distribution of the data; this yielded 11 attributes (8 quantitative and 3 categorical variables) to perform the final analyses (Table 1).

According to the PCA performed with the 11 variables, 73% of the total variance is explained by the first 3 components. The first component accounted for 46% of the variance, in which the attributes with the greatest contribution were height, leaf length, and width at the leaf base; the second component explained 19% of the variance, and the variables with the greatest contribution were tooth shape, number of leaves, and spine length; for the third component, the attribute with the greatest contribution was tooth length (Table 1).

Figure 4 shows that “Manso” plants are grouped to the left of the graph, along with the variables that characterize this group (plant, leaf, tooth, and spine size); “Amarillo” plants are located at the top right side, characterized by the curved tooth shape and leaf margin shape, while most “Prieto” plants were grouped at the bottom right side, identified by leaf shape. In other words, axis 1 separates “Manso” from the other 2 landraces, while axis 2 segregates “Prieto” from “Amarillo”.

Figure 4 Ordination plot of the Principal Component Analysis (PCA) of the 61 agave plants belonging to the 3 traditional landraces (m: “Manso”, a: “Amarillo”, p: “Prieto”), with regard to the 11 morphometric variables showing the greatest variation: SWB = spine width at the base, LWB = leaf width at the base, SL= spine length, LL = leaf length, NL = number of leaves; NT = number of teeth; PL = plant total length; TL = tooth length; SLL = leaf shape; ST = tooth shape; LMS = leaf margin shape. 

The hierarchical cluster analysis yielded a cophenetic relationship of 0.82, indicating that the clustering was well- supported. The dendrogram produced 3 groups that were confirmed through the PCA. The first group is composed mainly of “Manso”, with a classification certainty of 87%, which displayed a greater morphological variation. The second group clusters all “Amarillo” plants, characterized by curved teeth and a short spine; these agave plants are of intermediate size and, as shown in the dendrogram, are more similar to the “Prieto” agave. The latter is grouped in the cluster on the right, with a 58% certainty, characterized by ascending leaves, a smaller number of teeth, and a smaller plant size (Fig. 5).

Figure 5 Tree of morphological distances of the plants analyzed (r = 0.82). The clustering method used was the farthest neighbor approach. Values under each node correspond to the group; figures in red indicate the value that supports the group. Abbreviations: m = “Manso”, a = “Amarillo”, p = “Prieto”. 

Tukey’s test showed that 10 of the 11 variables displayed significant differences between landraces; 2 of these attributes were categorical and 8 were continuous (Table 2). “Manso” is the largest variety, with individuals measuring 2.82 m in height on average. This variety shows leaves and spines broader at the base and the longest teeth; these attributes are significantly different from those observed in “Prieto” and “Amarillo”, with average heights of 2.01 m and 1.92 m, respectively. “Amarillo” shows significant differences in the number of leaves (F = 15.861; p < 0.001) and spine length (F = 20.719; p < 0.001); this landrace has fewer leaves per rosette (18 on average), unlike the other varieties that have 34 or 37 leaves on average. “Amarillo” displays the shortest spines of the varieties studied, measuring an average of 5.47 cm. On the other hand, one of the characteristic attributes of the “Prieto” landrace is the rosette shape with densely packed lanceolate and ascending leaves, whereas the other varieties have loosely arranged lanceolate leaves.

Only 2 of the 9 variables were significantly different across the 3 varieties: one was leaf margin (F = 77.28; p < 0.001); in “Manso”, the leaf margin is straight (attribute coded with number 1), while “Amarillo” shows an undulated leaf margin and “Prieto”, semi-undulated leaf margin (coded with numbers 3 and 2, respectively). The number of teeth also differed significantly (F = 62.265; p < 0.001) across landraces, with “Manso” showing the largest number of teeth (60 on average), followed by “Amarillo” (48), and then “Prieto”, with the lowest (36).

Discussion

There are scarce records of Agave in Tlaxcala, Mexico. ^{Gentry (1982)}, reported no records of agave in Tlaxcala. Some studies have reported agave species and landraces used to produce “pulque” in Tlaxcala. ^{Mora-López et al. (2011)} reported, for 4 localities in 4 municipalities of Tlaxcala, Agave mapisaga “Sabililla”, A. mapisaga var. lisa and 4 landraces of A. salmiana var. salmiana. ^{Álvarez-Duarte et al. (2018)} recorded 9 landraces of A. salmiana subsp. salmiana for 9 localities of the municipalities of Nanacamilpa and Calpulalpan. Four landraces of Agave salmiana subsp. salmiana were reported in a locality of Nanacamilpa, Tlaxcala, (^{Álvarez-Ríos, Figueredo-Urbina et al., 2020}; ^{Ramírez-Manzano et al., 2020}). Besides, ^{Trejo et al. (2020)} recorded A. mapisaga var. mapisaga “Palmilla” and 8 landraces of A. salmiana subsp. salmiana for 10 localities in 9 municipalities of Tlaxcala. Finally, ^{Trejo et al. (2022)} reported, for the central zone of Nanacamilpa, Tlaxcala, A. americana “Cenizo”, A. mapisaga var. mapisaga “Palmilla”, A. salmiana subsp. tehuacanensis “Cimarrón”, 7 landraces of A. salmiana subsp. salmiana, and 2 unidentified landraces because they are no longer present in the area. It is worth noting that A. salmiana var. salmiana and A. salmiana subsp. salmiana can be considered synonyms, the former based ^{on Gentry (1982)} and the latter, on ^{García-Mendoza (2011)}.

Although SPT is a small locality, it harbors a high diversity of landraces used to produce “pulque”. Agave americana “Palma” and A. salmiana subsp. salmiana “Amarillo” and “Mahuahuac” have been reported only for SPT. All other landraces have been recorded at least in one other locality in Tlaxcala. This finding may reflect that indigenous populations growing agave under low cultivation intensity conserve a high landrace diversity, as suggested by some authors (^{Álvarez-Ríos, Pacheco-Torres et al., 2020}; ^{Reyes-Agüero et al., 2019}).

The names of “pulque” agave landraces are assigned by persons who are well-acquainted with these plants. Traditional identification is based on characteristics such as plant color, plant size, spine length, and leaf length. However, the same variety can be given various names in different zones or localities, hence the importance of analyzing the lexical marking since it can provide information on the cultural significance of that plant if its name does not change within a zone or in adjacent areas (^{Berlin, 1992}). A wide distribution of the “Manso” and “Colorado” landraces of Agave salmiana subsp. salmiana has been observed in the state of Tlaxcala (^{Álvarez-Duarte et al., 2018}; ^{Mora-López et al., 2011}; ^{Trejo et al., 2020}). “Manso” has been reported as the landrace most extensively used in the Central Mexican Plateau (^{Álvarez-Ríos, Figueredo-Urbina et al., 2020}; ^{Alfaro-Rojas, et al., 2007}; ^{Mora-López et al., 2011}). Some landraces thrive in the central region of the state, such as “Prieto” and “Matecón” landraces of Agave salmiana subsp. salmiana (^{Trejo et al., 2020}). A. mapisaga Trel. var. mapisaga “Penca larga” is grown in the southwest region of Tlaxcala (^{Trejo et al., 2020}).

The names used for the SPT landraces and those assigned in nearby regions are identified as the same species and are given the same common name; thus, the names indicate the cultural importance of these plants in the locality, in both the state of Tlaxcala and, actually in all central Mexico. It appears that landraces are developed from the selection of local resources, a phenomenon that has also been observed in other regions of Mexico (^{Álvarez- Duarte et al., 2018}; ^{Mora-López et al., 2011}; ^{Reyes-Agüero et al., 2019}). This fact highlights the importance of the knowledge of the traditional characteristics used in the identification of Agave landraces; this is highly relevant for biocultural conservation and underlines the risk of local extinction facing agave plants and the loss of knowledge on them.

As in other ethnobotanic studies about Agave (^{Álvarez-Duarte et al., 2018}; ^{Ramírez, 2007}; ^{Vázquez-García, 2018}), additional information was gathered from local experts (named “tlachiqueros”). The emigration of young generations has influenced the loss of traditional knowledge. The new generations are no longer interested in learning “tlachiquero” skills and prefer to emigrate to urban areas in search of better-paid job opportunities (^{Álvarez- Duarte et al., 2018}; ^{Vázquez-García, 2018}). In SPT, all respondents were consistent in naming and identifying agave landraces. However, only 1 respondent provided the name of landraces in both Nahuatl and Spanish. The loss of knowledge of the Nahuatl names may be due to the decrease in the number of speakers of this language and the little interest in learning “tlachiquero” skills. To date, there are scarce studies addressing the names and identification of agave plants from the traditional knowledge perspective (^{Colunga-GarcíaMarín and May-Pat, 1993}; ^{Ramírez-Manzano et al., 2020}; ^{Reyes-Agüero et al., 2019}).

In the state of Tlaxcala, the “Manso” landrace of Agave salmiana subsp. salmiana ranks first in the frequency of mentions, indicating the importance of this landrace for “pulque” production in the locality. The importance of this landrace has also been reported for a large portion of the “pulque”-producing zone in Mexico (Alfaro- Rojas et al., 2007; ^{Álvarez-Duarte et al., 2018}; ^{Álvarez-Ríos, Figueredo-Urbina et al., 2020}; ^{Mora-López et al., 2011}; ^{Trejo et al., 2020}). According to these reports, “Manso” is preferred because of its size, the flexibility of its leaves that facilitate handling, and, most of all, the high quality of its sweeter sap of 9-15 °Brix (Mora- López et al., 2011; ^{Ramírez-Manzano et al., 2020}; ^{Trejo et al., 2022}).

In SPT, “Amarillo” ranked second in frequency of mentions; this landrace is very scarce in the area and is rarely used, but draws attention for the relevance it still has among the local inhabitants. The respondents mentioned that fifty years ago, “Amarillo” was extensively used to extract “ixtle” (fibers of agave leaves) to make “cuartas” (ropes) used in religious festivities and to produce “pulque”. However, due to the use of synthetic ropes, the current “ixtle” extraction is nil, which led to the decline of “Amarillo” cultivation. The landrace ranking third in frequency of mentions was “Prieto”. ^{Trejo et al. (2020)} point out that this landrace is used to produce “pulque” mainly in the central region of Tlaxcala. It is likely that “Prieto” is synonymous with the “Negro” landrace also used for “pulque” production in the Estado de Mexico (^{Alfaro-Rojas et al., 2007}).

The analysis of morphological variables in this work recognized morphological clusters that correspond to the 3 traditional landraces mentioned most frequently and are the most abundant landraces in the locality studied. Agave salmiana subsp. salmiana, fostered by the selection and management practices of peasants. ^{Trejo et al. (2020)} observed morphological diversity within the subspecies salmiana; however, no such differentiation was evident at the genetic level using nuclear and chloroplast markers.These findings suggest an incipient domestication scenario.

Three of the 5 characters used by peasants to differentiate landraces (plant length, leaf length, and leaf width) have been significant in most studies addressing the morphological diversity of agave plants (Table 3). In contrast, characteristics such as plant color have been poorly analyzed and spine stiffness has not been evaluated in these studies. The elements used by peasants to identify landraces and the statistical analysis of the variability of morphological characters are useful for identifying the differences between landraces. Therefore, the 2 approaches complement each other to recognize diversity.

Several studies have mentioned that plant size, number of teeth, tooth size, and distance between teeth are associated with Agave domestication and suggest that artificial selection has caused this morphological trend (^{Alfaro-Rojas et al., 2007}; ^{Figueredo et al., 2021}; ^{Mora-López et al., 2011}). On the one hand, larger plants produce a greater volume of sap and are, therefore, associated with a higher sap production; on the other, smaller or more separated teeth facilitate manipulation (^{Figueredo et al., 2021}; ^{Mora-López et al., 2011}). This information partially coincides with the results of the present study since the preferred landrace by tlaquicheros for “pulque” production is “Manso”, a landrace characterized by intermediate plant size. “Manso” is neither a large-sized landrace nor the one with the highest sap production; it displays mid-sized lateral spines and flexible leaves (^{Alfaro-Rojas et al., 2007}; ^{Mora-López et al., 2011}; ^{Ramírez-Manzano et al., 2020}; ^{Trejo et al., 2020}).

The holistic or essentialist understanding of peasants to recognize their own landraces has not been addressed in detail. Understanding it may contribute to including additional characters to improve the description of the morphological diversity of landraces and the reasons for their selection. More detailed interviews with peasants would broaden our understanding of how they identify agave plants and better describe how agave cultivation is managed. Strategies should be developed aiming to gather more in-depth and detailed information from peasants, as the descriptions they have provided so far are just basic and general; the current approaches seem relatively ineffective as peasants do not provide detailed data and may even get impatient.

San Pedro Tlalcuapan is a small town that harbors a large number of landraces; however, most of these include just a few individuals, highlighting the urgent need to implement conservation strategies to prevent their disappearance. Such strategies shall promote the dissemination of these studies in the local communities and the use of all varieties, raise interest in the “pulque” production chain by the young generations, and promote the transfer of traditional knowledge and the learning of Nahuatl.

The qualitative morphological traits used by peasants to distinguish landraces are significant in most studies addressing the morphological diversity of agave plants. The two approaches complement each other and allow identifying biodiversity patterns. Furthermore, some traits used by peasants, such as the flexibility of the terminal spine, can be integrated into the traditional morphometric studies.