Relationship between populations of Ferocactus pilosus

The degree of affinity between F. pilosus populations is shown in Figure 1, which formed three main groups. The greatest similarity was between group 1 composed of the Mexican fire barrel cacti of populations I and V and group 2 of populations III and IV and these with group 3 of population II. This same answer would have been obtained in the tests of Manova and Tukey.

Figure 1 Dendrogram of the populations of the Highlands of Tamaulipas, which shows the existence of variables related to the morphometry of fruit and seed of Ferocactus pilosus (the name of the population is described in Table 1). 

However, based on the physical characteristics, populations II and III are similar in the type of climate, soil, geographical location, elevation and surface stoniness. Population IV, on the other hand, has a temperate subhumid climate defined by the higher elevation and lower stoniness and latitudinal geographical location located almost at the height of population II (Table 1).

These characteristics of these populations may have influenced the existence of significant differences in SN and SW of the Mexican fire barrel cactus (Table 2). The relationship that unites these three populations with populations I and V is found in a biotic-type response in which physiological maturity (^{Ayala-Cordero et al., 2004}) and population structure participate (^{Lara et al., 2016}), given the coincidence in age, size and number of stems of the Mexican fire barrel cacti at the time of fruit collection.

The distribution of F. pilosus in the Highlands of Tamaulipas extends to alluvial soils, where the greatest abundance and diversity of plant species is favored, but it is also located in areas of piedmont and abundant surface stoniness, with a predominance of limestone rocks that retain soil moisture. The results of ^{Tilman (1990}) highlight that the interaction between plants is traditionally analyzed in the aerial part, above the ground (energy translated into dry biomass) and the attention to the underground part (supply of water and nutrients), where the root acquires greater importance in environments of low productivity, is low.

The orographic-type shading that normally derives from the Sierra Madre Oriental favors the retention of water in the soil and its availability for plants (^{Salinas-Rodríguez, 2018}). It is known that water deficiency in a succulent stem tends to affect the storage organs, particularly between flowering and seed maturity (^{Tardieu, 2013}). In this sense, ^{Ayala-Cordero et al. (2004}) consider as a rule the fact that, in populations of Stenocereus beneckei located in environments restrictive of humidity, a variation in the number and weight of seeds.

The small size (< 1 mm) and the abundant production of seeds of F. pilosus increases its reserve in the soil and dispersion by anemochory (personal observation). Unfortunately, F. pilosus populations are exposed to predators and anthropogenic use of flower buds and fruits, an action that interrupts the reproductive cycle (^{Alanís-Flores and Velazco-Macías, 2008}) and decreases the potential distribution of populations.

This happens in Mexico for F. pilosus, which, despite occupying an extensive area (298 007 km²), has lost 16% of cover at the local level (^{Ballesteros-Barrera et al., 2017}). This requires the search for conservation actions, but importantly to the in-situ recovery of F. pilosus, reproduced under controlled conditions.

Statistical criteria for the selection of variables

The growth and development of the fruit involve internal and external plant factors, such as genetics, hormone content, position and competition of the fruit, as well as availability of water and nutrients, and temperature and luminosity (Fisher et al., 2018). This dependence may have been expressed through the correlation analysis, which reveals a strong positive association between PL vs ED (0.7422; p< 0.0001), PL vs RMx (0.7421; p< 0.0001) and PL vs FV (0.8780; p< 0.0001).

Despite the fact that there was a perfect positive association, although this same condition was maintained between ED vs FV (0.9262; p< 0.0001) and RMx vs FV (0.9262; p< 0.0001), it failed to influence the physical characteristics of the seed, such as SN vs SW (0.9999; p< 0.0001). Based on these results, the Manova was developed. Both SN (F= 3.49; p< 0.0085) as SW (F= 3.47; p< 0.0088) resulted in statistically significant differences when contrasting the populations of Ferocactus involved in the Manova.

In this model, 92% of the variance was explained by the first three characteristic roots, while the Manova procedure meets the multivariate F values and the criteria of Wilk’s Lambda= 0.8087 (F= 1.7; p< 0.0097), Pillai’s T= 0.2049 (F= 1.69; p< 0.0098) and Hotelling-Lawley’s T= 0.2201 (F= 1.7; p< 0.0103). That is, although the value of F varies for each statistic, each corresponding p value is less than α= 0.05; so, the null hypothesis of Manova is rejected and it is concluded that Ferocactus populations have an effect on SN and SW, which are discussed below.

Seed number

The largest number of Mexican fire barrel cactus seeds was recorded in fruits collected from populations II (674 ±191); III (657 ±221); and IV (643 ±246), with respect to the total mean (629), according to Tukey (α≤ 0.05). This answer proves the existence of competition in each population (^{Raisman and González, 2013}), in a restrictive semiarid environment where the requirements are similar but sufficient to form a significant number of seeds in the populations of Mexican fire barrel cactus.

Competition is often associated with the evolution of the strategies employed by the species, such as the nurse plants demanded by the Mexican fire barrel cactus in stages of initial growth and the multiplication of stems, which depends on a mature individual, while competitive ability is determined by phenotypic and genotypic attributes (^{Grime, 1977}), which change among barrel cactus populations. Individuals in populations II, III and IV of F. pilosus showed greater ability to channel their resources into reproductive structures.

A response that manifests itself in a greater variation in the number of seeds. The number of seeds found in F. pilosus is abundant, although lower than those formed by F. histrix, 2 200 seeds per fruit (^{Del Castillo, 1986}; ^{Loza et al., 2012}). Normally, seed abundance is accompanied by size reduction, observed in species with r-type reproductive strategy (fluctuating), such as that shown by F. pilosus: mass production of potential individuals (seeds) versus low probability of survival (seedlings).

With development, this species changes its strategy to type k (constant): bulky and long-lived individuals, but their growth is slow, and maturation is late. Therefore, the survival of F. pilosus populations depends on a small number of individuals or their low recruitment capacity (^{Morláns, 2004}). The number of seeds is extremely different between/within cacti, depending on age, number of flowers or fruit size (^{Santos-Díaz et al., 2010}), as well as the high dependence on pollinating agents, so the barrel cactus maintains its functionality and genetic variability at the expense of inbreeding (^{Stein et al., 2017}).

That is, the limited dispersal of F. pilosus seeds may have an impact on the local genetic structure in the short term (^{Grivet et al., 2009}), in view of the reduction of mothers and fathers, a progressive increase in genetic structure and levels of biparental inbreeding, which would explain the decrease in seeds produced, germination rates and vigor (^{Nora et al., 2011}) and low seedling establishment.

The number of F. pilosus seeds showed a behavior similar to that of isolated metapopulation species; that is, they show a simple allometric trajectory in the absence of competition (^{Torroba et al., 2013}), but intraspecific competitive interactions induce morphological variations in certain individuals (^{Schwinning and Kelly, 2013}), as occurred with SN.

Another cause of the specific variation of the seed of F. pilosus is the characteristics that increase the probability of survival in a given environment (^{Martínez et al., 2016}), without underestimating the existence of some type of pressure of different nature at the present time. ^{Larios et al. (2014}) established that the critical resources with which competition operates vary between environments and are increased by resources such as soil nutrients and water in dry environments.

Unfortunately, the ultimate effect of intraspecific competition manifests itself in future generations (^{Valeria et al., 2017}); therefore, this type of competition would tend to regulate the size of the populations of Mexican fire barrel cactus. Number and weight of cactus seed also depend on the growing environment, age and adaptation of the reproductive system (^{Ayala-Cordero et al., 2004}) and pollinators (^{Valiente-Banuet, 2002}).

Seed weight

In the seed weight (SW) of Ferocactus pilosus, populations II, III and IV were outstanding, with 0.92 ±0.27, 0.9 ±0.3 and 0.88 ±0.34 mg, each, statistically different according to the Tukey test (α≤ 0.05). Some explanations around this response could be due to a similarity in the best use of low soil moisture, nutritional reserves, solar radiation, temperature, age of the plant, prevailing between flowering and seed formation and maturity (^{Ayala-Cordero et al., 2004}; ^{Flores and Jurado, 2011}) in growing environments with some degree of similarity between the three populations of barrel cactus, with repercussions on seed weight.

The weak to moderate positive correlation that existed between SW with PL, ED, RMx, RMn and FV would indicate that the weight gain of the seed does not depend on the size of the fruit, which is highly variable and therefore an undesirable attribute for F. Pilosus. Among these, the low contribution of PL in SW would also have been found in F. robustus and Pterocereus gaumeri (^{Méndez et al., 2005}; ^{González and Navarro, 2011}), in contrast to the positive response obtained in F. cylindraceus, F. wizlizenii and Echinocactus platyacanthus (^{McInthosh, 2002}; ^{Díaz et al., 2008}).

The seeds of the barrel cacti of populations II, III and IV, due to their greater weight, could have an advantage in terms of survival, viability, germination and establishment, with respect to the seeds of populations I and V of F. pilosus; nevertheless, the weight of the seed decreases as its number in the fruit increases. In the same sense, the seeds of the Mexican fire barrel cactus with variable weight in a population would be expected to exhibit memory of hydration and alternating germination (^{Contreras-Quiróz et al., 2016a}, ^2016b).

The variation in weight of seeds of the same species responds to a differential biological capacity (^{Sánchez-Salas et al., 2006}). An ideal advantage of large seeds, such as those of barrel cactus populations II, III and IV, is to have higher nutritional reserves and despite germinating slowly, they have high germination and emergence rates (^{Brown et al., 2003}). In reality, the Mexican fire barrel cactus has a high mortality rate during germination and establishment in natural conditions, which motivates the transplantation of individuals reproduced ex situ.