Study area

The study was carried out during 2016 in Santa María Zoquitlán, municipality of the same name, which is a community of the Central Valleys of Oaxaca, located at 16°34 'N and 96°21' W, at an average altitude of 1 080 m. The predominant vegetation is the low deciduous forest, where B. simplex, an endemic species of the Tehuantepec River basin, is distributed. The genera that share the habitat are Cephalocereus, Cercidium, Lysiloma, Ceiba, Neobuxbaumia, Escontria, Acacia, Beaucarnea, among others. The predominant climate is the type BSo (h ') w "(w) i' g (^{Acosta et al., 2003}).

Sampling method and data collection

Selected samplings were carried out with the aim of identifying trees of the species of interest with outstanding phenotypic characteristics, of healthy appearance and that also had fruits, taking this as the main criterion for their selection. During October-December 2015, 27 trees were identified in the area called El Paredón, located in an altitudinal interval of 1 078-1 406 m; the trees were marked with progressive numbers and were taken as a reference point to establish 23 sample units (SU) of 100 m².

At each sampling site all the present species were counted and with the support of a Fiber-glass measuring tape of 1.52 m and a 5 m measuring tape (Chain brand, MGA 5020 model) the morphological variables corresponding to the collar diameter were measured and recorded (DP, cm), total height (AT, m) and crown diameter (DC, m) with which the crown area was calculated (AC= 1/4 × πDC², m²).

In the sample units established with a Garmin DriveTM GPS, the site variables (exposure, altitude (m) and slope (%) were recorded with a Suunto Finland clinometer Pm5/36pc.

From each of the selected Bursera specimens, part of the fruits presented was collected, which were placed in paper bags and moved to the facilities of the Instituto Tecnológico del Valle de Oaxaca (ITVO) (Technological Institute of the Valley of Oaxaca) (ITVO).

A random sample of 50 fruits of each tree was taken, in which they were measured with a Pretul vernier of 0.1 cm of precision, their polar diameter (Dp, mm) and their equatorial diameter (DE, mm); the form coefficient (FC = DE / Dp) was obtained and weighed (g) individually with an Ohaus Adventurer™ analytical balance with a maximum capacity of 210 g and precision of 0.1 g. to obtain the number of fruits kg^-1. Subsequently, for drying, they were placed under solar radiation for four days; next, the seed was separated from the tissues of the fruit, to then measure its polar and equatorial diameter; finally, they were weighed to obtain the number of seeds kg^-1.

With the morphological variables obtained, structural indexes were calculated and frequency histograms were made, using commercial diameter categories of 5 cm and height categories at intervals ≤1.5, (1.5 to 3), (3.1 to 4.5) and (4.6 to 6), > 6 in meters. Three structural assessment indices were determined (^{Zarco et al., 2010}; ^{Santiago, 2013}):

Where:

IVI = importance value index

DMr = Relative dominance or relative basal area (ABr)

Dr = Relative density

Fr = Relative frequency

IVF= Forest value index

DPr = Relative collar diameter (cm)

Ar = Relative height (m)

ACr = Relative crown area (m²)

IC= Complexity index

s = Number of species

d = Density of trees larger than 2.5 cm DP

b = Basal area (m²)

h = Height (m) of the three dominant trees

A= Sampling area (m²)

Data handling and analysis

In the fruits, seeds and trees data, the assumptions of normality and homogeneity of variances were verified by the Shapiro-Wilk and Bartlett tests. The variables that did meet the assumptions of normality and homogeneity of variances were subjected to analysis of variance by a completely randomized design and comparisons of means were made by the Duncan test (α = 0.05), in which the classification variable was used tree of origin, exposure, altitude and slope; for these last two, three categories were formed individually.

For the analysis routine, the SAS computer program (2004) was used, with the GLM procedure. In the case of the variables that did not comply with the assumptions of normality and homogeneity of variances, they were analyzed by means of the Kruskal Wallis test with the NPAR1WAY Wilcoxon procedure, by using SAS statistical package (^{SAS, 2004}).

Vertical and horizontal structure

The differences in the size of trees of the same species located in different exposure sites can be attributed to the fact that the magnitude of growth is a function of the potentials defined by the genotype, the environment and the interaction of both. ^{Ramírez and Rodríguez (2004)}, comment that the difference in the growth of trees is attributed to the fact that in the northern hemisphere the southern exposure receives the greatest amount of solar radiation. In this sense, ^{Ortiz-Pulido and Pavón (2010)} mention that the exposure of the site and the amount of solar radiation that affects influences the sexual differentiation of the trees of Bursera, since the trees in sites facing east, have more frequency those of male sex and in the groups of trees facing west, female sex is more frequent.

^{Sánchez and López-Mata (2003)} state that altitude is a physical factor that influences climatic conditions through variables such as temperature and precipitation, which, in turn, have a direct effect on plant growth and are correlated specifically in each site. ^{Martínez et al. (2013)} mention that relationships have been found between the growth of conifers and the factors of soil depth, altitude, slope and exposure. ^{Rubio et al. (2011)}, when studying Quercus vegetation, attributed the observed differences to the plant successional state and that the early successional species produce smaller seeds.

The maximum heights in the different sample units (SU) were 6.60 m; for descriptive purposes of the vegetation, two strata were identified: inferior and superior. The lower stratum is constituted, mainly, by individuals ≤1.5 m, reunited, in its majority, in the western exposure (1 057 trees ha^-1) and 326 individuals ha^-1 greater than 1.5 m. The lowest density of arboreal specimens in the lower stratum was counted (209 individuals ha^-1), in the sites located in the southwest exposure. The trees of B. simplex make up the upper layer with heights ≥ 2 m (Figure 1a).

Densidad (individuos ha ^-1 ) = Density (individuals ha^-1); Exposición = Exposure

Figure 1 Frequency of individuals according to their categories of height (a) and diameter (b) at different exposures. 

^{Gutiérrez et al. (2011)}, in the vegetation of low deciduous forest in three localities of the state of Yucatán, recorded similar data, which they obtained from groupings with individuals of 1.5 to 4.4 m. ^{Dzib-Castillo et al. (2014)}, when referring to the vegetation in the communities of Pomuch and San José Carpizo in the state of Campeche, calculated heights of 6.3 m in low deciduous forest and diameters of 6.9 cm. On the other hand, in a low deciduous forest in Xmatkuil, Yucatán State. ^{Gutiérrez and Zamora (2012)} measured individuals from 4 to 12 m in height and ^{Castillo et al. (2007)}, in the vegetation of Veracruz State, described three arboreal strata from 1 to 7 m.

In the present work, a greater density of individuals with diameters <2.5 cm in the horizontal structure was quantified in all exposures, followed by diameters greater than 2.5 and less than 7.5 cm, very related to the heights that were more pronounced in the west exposure (Figure 1b); this is explained by the frequency of individuals with higher diameters, which decreases as they reach greater dimensions, which means that the smaller diameter classes will always be more abundant.

The structural indexes (IVI, IVF, IC) showed significant statistical differences (p≤ 0.0001), in their effect on the morphological features of the fruits (Table 1). These differences are mainly due to the fact that the structure is determined by morphological characteristics of trees located in a specific area, as well as by density and competition, which implies dynamic relationships between individuals either to compete for nutrients or to obtain better development of its structures. Therefore, not always greater complexity will represent greater stability of the ecosystem; in addition, the spatial arrangement of individuals determines the particular morphology of the plant (^{Del Río et al., 2003}; ^{Perreta y Vegetti, 2005}). ^{Alvis (2009)} and ^{Zarco et al. (2010)} indicate that the IVI reveals the ecological weight of a species within a population.

Fruits and seeds characteristics

The fruits and seeds collected from 27 Bursera simplex trees showed significant statistical differences in weight, form coefficient (FC) and number of seeds kg^-1 (p ≤ 0.0001), depending on the tree of origin. The greatest heterogeneity (VC> 65 %) was shown in the weight of the seed and number of seeds kg^-1; the lowest variation was recorded by the FC of the fruit and the FC of the seeds (VC = 7.42 and 13.36 %, respectively) (Table 2).

From tree 5, fruits of larger size and weight were selected (0.42 ± 0.09 g), so that the number of fruits kg^-1 is low (2456 ± 571.46); these magnitudes were significantly different (Duncan, 0.05) from the sizes and weights of the fruits from trees 3, 9 and 16, which weighed 0.18 ± 0.03 g and the number of fruits kg^-1 was 5 743.7 ± 1 138.21, 5 566 ± 998.51 and 5 743.7 ± 1 138.21, respectively.

Given the relation equatorial diameter-polar diameter, the FC in fruit in trees 1, 15, 19 and 23 was higher (0.81 ± 0.06 cm) as in tree 10 (0.81 ± 0.05 cm); this implies fruits of almost spherical shape.

The highest FC in the seeds was recorded in the fruits of tree 25 (0.79 ± 0.01 cm), which was statistically different (p = 0.0001). The tree 21 presented the seeds of greater weight (0.08 ± 0.13), which produce 16 179 ± 5 036 seeds kg^-1, while those of tree 3 were smaller, which would have 35 489 ± 10 209 seeds kg^-1 (Table 3).

Possibly the morphological variations in fruits and seeds of the different trees result from Bursera simplex's own reproduction type, since its dioic condition promotes cross-pollination, and, therefore, greater diversity in the offspring (^{Edwards y Sharitz 2000}; ^{Enríquez et al., 2004}; ^{Rzedowski y Calderón 2006}; ^{Magallán et al., 2009}). The morphological variability observed in fruits and seeds could have a remarkable genetic component besides the environmental one (^{Gómez et al., 2010}).

On the other hand, ^{Guzmán and Cruz (2014)} report that the growth of the fruit is related to the available humidity, since with high temperatures with little humidity, the fruits become smaller, which varies according to the position they have on the tree; the large fruits develop in the upper part of the crown due to the greater availability of sugars and proteins produced by the photosynthetic area, in comparison with those of the lower part of the crown. Based on the above, ^{Rubio et al. (2011)} mention that the variation in the size of the seeds is important because it provides diverse capacities to settle in sites with different conditions.

^{Gonzáles and Cruz (1995)} establish that fruits with FC close to the unit have homogeneous diameters (polar and equatorial) and have a spherical shape, while values that diverge from it describe elongated fruits. Site conditions such as altitude, exposure and the slope of the terrain where the tree grows influence the magnitude of its growth and these in turn influence the morphological characteristics of fruits and seeds (p ≤ 0.01). Fruits of greater weight 0.34 ± 0.07 g were harvested from trees located in sites with southern exposure; seeds collected from trees growing in sites with southwestern exposure weighed 0.06 ± 0.07 g; in contrast, those of trees in north and northeast exposures weighed 0.03 g.

The FC of fruits and seeds reflects significant statistical differences (p ≤ 0.0002); therefore, it is confirmed that the fruits with the highest FC were those harvested from trees in sites with east exposure with 0.81 ± 0.06 cm, which contrasts with the seeds from trees in southern exposure whose CF was 0.72 ± 0.10 cm. On the other hand, the seeds collected from trees in sites in southwest exposure had a higher FC (p = 0.72 ± 0.09 cm). The amount of fruits and seeds kg^-1 was statistically different (p = 0.0001), because the seeds collected from trees located in northeast exposure accumulated 4 235 ± 1 338 kg^-1 fruits, higher than the southern exposure, 3 054 ± 787 kg^-1 fruits; similar case occurs in tree seeds in northeast exposure, which gave a total of 30 124 ± 8 746 (Table 4).

The slopes of the evaluated sites showed significant differences (p ≤ 0.001); from trees in sites with 25 and 40 %, the fruits of greater size were collected, with an average of 0.31 g and the production of fruits kg^-1 is favorable in slopes> 40 %, since they sum up an average of 4 104, with a FC equal to 0.71 ± 0.008 cm. From trees located in places with small slopes, fruits that tend to be rounded were taken, unlike the fruits of sites with steeper slopes, which outline an elongated shape.

The same behavior was not observed does with the seeds collected from trees located in slopes <25 %, since between the slopes 25 and 40 %, they reached a greater weight (0.06 g), with shape form of 0.71, while the seeds from trees located at slopes> 40 % were smaller but more abundant (27 040 seeds kg^-1) in contrast to the seeds taken from trees located at sites on slopes <25 % that were larger, and that gave a total of 16 789 seeds kg^-1 (Table 4).

The altitude of the sites in which the trees grow was also a factor with significant effect (p≤0.01), since in the altitudinal interval between 1 178 and 1 278 m, the weight of the fruits was 0.36 g, with the lowest FC (0.74); these data contrast with those occurring at altitudes ≤1 178 m, whose FC was 0.76, at this same altitude the fruit production was 4 441.

The seed had a similar response with 0.04 g at altitudes> 1 178 m; however, seed production was higher at altitudes ≤1 178 m (29 259.4) with form coefficients of 0.72 cm at altitudes ≤1 278 m (Table 4).

The morphological differences of the fruits and seeds in relation to the characteristics of the site (exposure, slope and altitude) can be understood as they are factors that influence the solar radiation that affects the plant, temperature and environmental humidity, conditions that affect the primary productivity of the species. In their study conducted in Quercus, ^{Márquez et al. (2005)} stated that the size differences are due to biological and physical factors, in addition to the size of the seeds in a species can vary between populations or between individuals, either for genetic reasons or for the life history of each plant (^{Griffiths et al., 2009}; ^{García et al., 2015}).