Study zone

It comprised the entire state of Nayarit, which is located between 23º 05’ and 20º 36’ north latitude and 103 º 43’ and 105º 46’ west longitude. It has an extension of 27 857 km², for which it occupies the 23^rd place nationally since it represents 1.4% of the total area of the country (^{INEGI, 2015}).

Regionalization and climatic characteristics

Due to its climatic and altitudinal characteristics, the study area was divided into three polygons: Costa Region (composed of the municipalities of Tecuala, Acaponeta, Rosamorada, Ruiz, Tuxpan, Santiago Ixcuintla, San Blas, Compostela and Bahia de Banderas) where a prevailing subhumid warm climate of the Aw type with an annual average temperature higher than 22 ºC and a temperature of the coldest month greater than 18 ºC and an altitudinal range that goes from 0 to 300 meters above sea level. Valles Region (which includes the municipalities of Tepic, Xalisco, Santa Maria del Oro, Jala, San Pedro Lagunillas, Ahuacatlan, Ixtlan del Río and Amatlan de Cañas) with a subhumid semi-warm climate of the type (A)C(w) with an average annual temperature greater than 18 ºC, temperature of the coldest month less than 18 ºC and an altitudinal range from 500 to 1 600 m.

Sierra Region (the municipalities that comprise it are Huajicori, La Yesca and Del Nayar) with a temperate subhumid climate of type C(w) with an annual average temperature between 12 ºC and 18 ºC, the coldest month between -3 ºC and 18 ºC and an altitudinal range between 900 and 1 200 m with one with a drier month precipitation of less than 60 mm, summer rains with precipitation/temperature index between 43.2 and 55.3 and winter rainfall percentage of 5% to 10.2% of the annual total (^{García, 1964}).

Genetic material

Seven primary races of native maize with current prevalence in the State of Nayarit were explored: Tabloncillo (42 collections), Tuxpeño (30), Elotero de Sinaloa (10), Blando de Sonora (3), Western Elotes (4), Bofo ( 2) and Vandeño (1), and the 11 secondary races or subraces that show influence or genetic infiltration of a second race: Tuxpeño x Tabloncillo (17), Tabloncillo x Tuxpeño (15), Western Elotes x Tabloncillo (2), Tabloncillo x Olotillo (1), Tabloncillo x Blando de Sonora (1), Western Elotes x Elotero de Sinaloa (1), Vandeño x Tabloncillo (3), Reventador x Tabloncillo (3), Olotillo x Tabloncillo (1) and Harinoso de Ocho x Tabloncillo (1).

Sites and number of collections

The harvest period of native maize ranged from 2008 to 2012, in collaboration with the National Plant Genetic Resources System (SINAREFI) of SNICS- SAGARPA, the National Commission for the Knowledge and Use of Biodiversity (CONABIO)-SEMARNAT, Foundation Produce Nayarit, AC and Municipal Presidencies. The prevailing state geographical composition was covered: Valles Region (RV), Sierra Region (RS) and Coast Region (RC), each of the collection sites being georeferenced. In 2008, 68 collections in eight municipalities: Amatlan de Cañas (4) (RV), Ixtlan del Rio (8) (RS and RV), Ahuacatlan (12) (RV), Santa Maria del Oro (9) (RV), Del Nayar (3) (RS), La Yesca (21) (RS), Tecuala (6) (RC) and Acaponeta (5) (RC). During 2010, 42 collections in eight municipalities: Seven in the Coast Region (RC): Banderas Bay (16), Compostela 85), San Blas (2), Santiago Ixcuintla (2), Tuxpan (6), Rosamorada (3) and Tecuala (6). Of the Sierra (RS) Huajicori Region (2) and in 2012, 56 collections made in the municipality of Santiago Ixcuintla (RV and RS). A total of 166 collections were made, most of them performed in RC with 107, RV with 33 and RS with 26 collections. The total number of localities sampled was 58 (Figure 1).

Figure 1 Geographic distribution of the collections, according to primary and secondary races and region in the state of Nayarit. 

Collection methodology

In each of the agricultural regions of the state (Costa, Sierra, Valles) were conducted tours for which work groups were organized, the collection periods were based on the harvest and post-harvest periods. The work scheme was based on the methodology and passport sheet suggested by the CONABIO (^{Ortega et al., 2013}) through collection routes by municipality, whose collectors gathered direct information from the producer or indirectly from a relative or acquaintance, by location. For the development of this study, the following aspects were considered for the field work and the collection of native corn: a) the number of collections per sampling site, was determined according to the diversity prevailing in the germplasm planted and harvested; by presence of representativeness and inter-racial variation, uses, time of planting by the farmer and variation in the environment in which it was sown; and b) sample size: in each collection, a minimum of 20 and up to 50 cobs of corn representative of the genetic diversity of each native maize existing in the community was programmed; in cases where the producer had already shelled, seed (4.5 kg as sample size) that were included in the present study as grain characteristics was collected. The cobs collected were applied with Deltamethrin (10 ml L^-1 of water), to eliminate and prevent the infestation of insects that damage the grain, being dried at room temperature until reaching a content of 10% moisture in grain. In most cases a representative sample of uncovered cobs was retained for identification and photography.

Characterization of the collections and taxonomic determination of races

The following concepts were considered in the passport sheet: primary race, secondary race, collector data, date of collection, catalog data, state, municipality, location, georeferencing of the collection site, determinator, name and data of the farmer, characteristics of the sample collected, uses, destination of production, characteristics of the sample collected from cob and gram, as well as agronomic management. Quantitative data were considered based on the morphological characteristics of the total number of cobs reported per collection and a representative and random subsample of grains (n= 5). The taxonomic identification of the maize races found was carried out with the support of participating experts, and based on the available literature (^{Sánchez et al., 2000}; ^{Ron et al., 2006}; ^{Ruiz et al., 2008}).

Database

The information on the passport sheet containing the data for the characterization of cobs and grain was protected in an information system for the CONABIO Biotic 4.5 database (^{Ortega et al., 2013}). A second database was created generated with the environmental variables prevailing in the collection sites. The final database was constituted by integrating the two previously described.

Plant variables

For the purposes of the present work, only the following variables of cob and grain (in cm and ml) were considered: length of cob (LM), cob diameter (DM), number of rows (NH), number grain per row (GH), grain width (AG), grain length (LG), grain thickness (GG), grain volume (VG).

Environmental variables

The environmental characterization was carried out with the data from the climate information system interpolated by the INIFAP (^{Ruiz et al., 2018}) at a resolution of 90 m. The climatic variables obtained from the system for each sampling site were: annual precipitation (PAA), average annual minimum temperature (TIA), average annual maximum temperature (TXA), annual average potential evapotranspiration (ETPA) and annual humidity index (IHA) which is the relationship between precipitation and potential evapotranspiration (PAA/ETPA).

Statistical analysis

A multivariate analysis of main components was carried out from the matrix of correlations between the attributes of cob, grain and the environment. For the selection of the components, the criteria were used: the first two that characterized the variation of at least 80% of the observations and that the typical values were >0.3 (^{Johnson and Wichern, 2007}). To determine the similarity of the races by the geographical origin of the samples, a multivariate cluster analysis was performed for the samples with racial origin Tabloncillo and Tuxpeño (^{Mohammadi and Prasanna, 2003}). The analysis was done with the statistical program Minitab version 17 (State University of Pennsylvania, United States of America).

Characteristics of cob

With the averages of the selected cob variables (to be used as descriptors and yield components), the collections were subjected to a principal component analysis (CP), the first two components together explained 81.5% of the observed phenotypic variation, where the first component (CP1) with a characteristic value of 1.73, explained, 43.3% of the total variance; the second component (CP2), whose characteristic value was 1.52, referred 38.2%. The variables with the greatest influence in CP1 were cob length (LM) and grains per row (GH). In CP2, the variables with the greatest discriminatory effect were cob diameter (DM) and number of rows (NH) (Table 1).

Based on the dispersion of the 166 collections made through CP1 and CP2, the variables LM and GH were identified as criteria for grouping the primary and secondary races within the ethnogenetic pattern of the Tabloncillo race: similarly, the DM and NH characteristics allowed to identify those primary and secondary races convergent with the ethnogenetic pattern of the Tuxpeño race (Figure 2). This coincides with the descriptors used to characterize these races (^{Rocandio et al., 2014}), which identified LM, NH, as appropriate variables for the racial characterization of native maize. The Tabloncillo cobs are identified by having a longer length and consequently more grain per row, while the cobs of the Tuxpeño race, although they are shorter, have a larger diameter and consequently a greater number of rows than Tabloncillo (^{Ortega et al., 2013}).

Figure 2 Dispersion of 166 maize collections made in Nayarit, based on the first two main components, on cob characteristics. A) Tabloncillo group; and B) Tuxpeño group. 

Grain characteristics

The first two main components with the averages of the grain variables selected, explained 97.9% of the phenotypic variation observed, where the first component (CP1) with a characteristic value of 3.81, explained 95.3% of the total variance and the second component (CP2), whose characteristic value was 0.1, showed 2.6%. Of the grain variables analyzed, the original variables with the greatest influence in CP1 were grain width (AG) and grain length (LG), while in CP2 the original variables of greater importance were grain thickness (GG) and grain volume (VG).

As the variables, AG and LG used as criteria of grouping of all the explored genetic material, they allowed to identify the primary and secondary races grouped within the ethnogenetic pattern of the Tabloncillo race; based on said grain characteristics. In the same way, the characteristic VG allowed to identify those primary and secondary races convergent with the ethnogenetic pattern of the Tuxpeño race, while the grain thickness (GG) was located in both varietal patterns (Figure 3). ^{Rocandio et al. (2014)} when describing seven maize races of High Valley, when making their AG morphological and agronomic identification, they mentioned LG, GG as appropriate variables to carry out their racial description. Similarly, ^{Rincón et al. (2010)} when studying the diversity and distribution of native maize, they use grain characteristics as descriptors of racial identity.

Figure 3 Dispersion of 166 maize collections in Nayarit, based on the first two main components, on grain characteristics. A) Tuxpeño group; and B) Tabloncillo group. 

Environmental variables

With the five environmental variables selected, the racial genotypes were subjected to the principal component analysis (CP). The first two components explained 67.8 and 25.9% of the total variation, respectively and an accumulated variance of 93.7%. In CP1 the annual precipitation (PAA), the minimum annual average temperature (TIA) and the annual humidity index (IHA); contributed mostly to the variation explained by this component. Similarly, in CP2, the most important characteristics were the mean annual maximum temperature (TXA), and the annual mean potential evapotranspiration (ETPA) (Table 1). In this way, the IHA and PAA variables were the criteria for grouping the primary and secondary races within the ethnogenetic pattern of the Tabloncillo race. The ETPA characteristic allowed to identify those primary and secondary races convergent with the ethnogenetic pattern of the Tuxpeño race, while the grain thickness was located in both varietal patterns (Figure 4). From the above, it is inferred that the environmental conditions where the Tabloncillo varietal pattern prevails, due to its early and intermediate vegetative cycle, archetype of lower leaf density and low-intermediate size are more dependent on the available humidity (PAA and IHA) than on the temperature (^{Ron et al., 2006}; ^{Ruiz et al., 2008}).

Figure 4 Dispersion of 138 maize collections carried out in Nayarit, based on the first two main components, on climatic characteristics. A) Tuxpeño group; and B) Tabloncillo group. 

The Tuxpeño varietal pattern, due to having a late vegetative cycle, archetype of abundant leaf density and tall size, have higher thermoenergetic requirements (TXA) and consequently require higher potential evapotranspiration (ETPA) (^{Ruiz et al., 2008}). The previous, also in consequence, is reflected in a greater yield of grain in the Tuxpeño race (2.9 t ha^-1) than in the Tabloncillo (2.5 t ha^-1) (^{Vázquez, et al., 2003}).

The wide distribution of the ethnogenetic patterns of Tabloncillo and Tuxpeño is widely supported by the presence of ecophysiological indicators in the different maize races, which show that this species currently has a wide range of responses to the environment, which promotes an inter-racial variation significant, that validates the influence of climate diversity on the cultivation of corn and are a sample of the genotype-environment interaction that has been occurring from the dispersion of corn throughout the Mexican territory (^{Ruiz et al., 2013}).

Ecological niches of ethnogenetic patterns Tabloncillo and Tuxpeño

The great capacity of response that the races have to the diversity of the environments where they are cultivated (^{Ruiz et al., 2008}) have favored the proliferation of ecological niches possessing an environmental potential (^{Muñoz, 2003}). The favorable environmental conditions prevailing in Nayarit, over time have allowed the potential distribution of the Tabloncillo and Tuxpeño maize races and their genetic combinations as secondary races, to become current distribution (^{Ruiz et al., 2013}).

In this way, Tabloncillo was found in the ethnogenetic pattern, at least the presence of four ecological niches given its greater environmental similarity: Niche 1: Acaponeta - Rosamorada and Niche; 2: Ruiz-Tecuala because both are located in the coastal region of the state. Niche 3: San Pedro Lagunillas-Tepic-Jalisco-Santiago Ixcuintla, and Niche; and 4: Ixtlan-Santa Maria del Oro, for its location within the Valles and Sierra regions (Figure 5A). For the ethnogenetic pattern Tuxpeño, the ecological niches determined in the present study were three, which presented a greater amplitude of distribution Niche 1: Ixtlan del Río-Rosamorada-Ahuacatlan-Santa María del Oro, in the Sierra and Valles regions; Niche; 2: Ruiz-Santiago Ixcuintla-La Yesca, in the Sierra region; Niche; and 3: from Nayar-Jala-San Pedro Lagunillas-Tecuala-Tepic-Xalisco-San Blas, with greater presence in the Valles region, followed by the Costa and Sierra region (Figure 5B).

Figure 5 Dendrograms of the state distribution of ethnogenetic patterns of races A) Tabloncillo (TBO); and B) Tuxpeño (TXP). 

This can be attributed to the fact that the great distribution of the agroclimatic conditions where corn is grown has led this species, through acclimation and adaptation processes, to express its genetic and phenotypic plasticity, above all due to the presence of natural and induced hybridization; which has led to a greater presence of interracial genetic diversity through environments (^{Ruiz et al., 2013}).

Thus, the highest prevalence of four ecological niches within the Tabloncillo ethnological pattern, indicated the presence of genetic flows between the breed Tuxpeño, Elotes Occidentales, Elotero de Sinaloa, Olotillo, Vandeño, Blando de Sonora and Harinoso de Ocho and Vandeño. Within the Tuxpeño varietal pattern, its genetic infiltration with Tabloncillo predominated.

This movement of germplasm has allowed the permanence and existence of these races through interracial genetic infiltration through the regions where the native maize is cultivated. This genetic flow can be attributed to the absence of gametophytic incompatibility in the Tabloncillo and Tuxpeño breeds because they are recessive homozygotes in the presence of alleles of incompatibility (^{Padilla et al., 2012}).

This has allowed the native germplasm to continue covering mainly the needs of the ethnic groups and rural population that continue using and conserving these native maizes for their agronomic-food attributes (^{Hernández, 1970}; ^{Muñoz, 2003}; ^{Ron et al., 2006}; ^{Rincón et al., 2010}; ^{Vidal et al., 2017}).