Introduction
Seed germination and seedling establishment are critical stages in a plant's life cycle (Fenner, Reference Fenner1985; Pérez-Sánchez et al., Reference Pérez-Sánchez, Jurado, Chapa-Vargas and Flores2011; Arroyo-Cosultchi et al., Reference Arroyo-Cosultchi, Golubov and Mandujano2016). Therefore, the development of strategies that maximize survival throughout these phases are of vital importance for the perpetuation of plant species. One such strategy is seed heteromorphism (Matilla et al., Reference Matilla, Gallardo and Puga-Hermida2005), or the formation of different seed morphs from the same individual (Venable, Reference Venable1985; Smith et al., Reference Smith, Wang, Msanga and Vozzo2002). This strategy has been described for 200 species and 18 families of angiosperms, with Asteraceae, Amaranthaceae and Brassicaceae as the most represented (Matilla et al., Reference Matilla, Gallardo and Puga-Hermida2005).
Variations between seed morphs can be described in terms of size, shape, colour and seed mass, as well as seed-coat sculpturing, water permeability, cuticular structure, gene structure and embryo size (Smith et al., Reference Smith, Wang, Msanga and Vozzo2002; Baskin and Baskin, Reference Baskin and Baskin2014; Leverett and Jolls, Reference Leverett and Jolls2014). Other variations between seed morphs also occur in seed behaviour without visible morphological differences (Venable, Reference Venable1985). This type of seed heteromorphism, termed cryptic, has been observed in various species (Leverett and Jolls, Reference Leverett and Jolls2014; Liyanage et al., Reference Liyanage, Ayre and Ooi2016), and is a common, yet rarely documented phenomenon (Venable, Reference Venable1985; Leverett and Jolls, Reference Leverett and Jolls2014). The ecological advantages of seed heteromorphism include differences between morphs in competitive ability, dispersion, dormancy, germination time, vulnerability to predators and seedling growth (Matilla et al., Reference Matilla, Gallardo and Puga-Hermida2005; Lu et al., Reference Lu, Tan, Baskin and Baskin2010; Leverett and Jolls, Reference Leverett and Jolls2014). However, understanding the selective forces that favour the evolution and maintenance of seed heteromorphism through evolutionary time remains a key question in studies dealing with heteromorphism; these forces probably depend on habitat and the mechanisms controlling population dynamics in the communities in which heteromorphism occurs (Liyanage et al., Reference Liyanage, Ayre and Ooi2016).
In the field we observed two seed morphs in Leptocereus scopulophilus Areces, an endemic threatened Cuban cactus. One of these morphs shows an apparently natural scarification of the seed coat, which may indicate a zone of lower mechanical strength favourable for germination. Seed heteromorphism in Cactaceae is barely documented. It has been reported for both seed size and mass in Astrophytum myriostigma Lem. (Sánchez-Salas et al., Reference Sánchez-Salas, Flores and Martínez-García2006) and Stenocereus beneckei (Engelm.) Buxb. (Ayala-Cordero et al., Reference Ayala-Cordero, Terrazas, López-Mata and Trejo2004). In these species, seed size implies differences in germination capacity and speed (Ayala-Cordero et al., Reference Ayala-Cordero, Terrazas, López-Mata and Trejo2004; Sánchez-Salas et al., Reference Sánchez-Salas, Flores and Martínez-García2006). However, to date no study on seed heteromorphism has reported apparently scarified seeds, and as such the possible implications of this trait in the germination of the species that possess it are not known.
In this study, the two seed morphs of L. scopulophilus are characterized for the first time; germination is described for each morph and performance is compared between the two morphs. We hypothesize that there are differences in germination capacity between the two seed morphs. Moreover, this study helps to clarify the implications of seed heteromorphism in the germination process of this cactus, which provide essential elements to consider in the conservation strategies of L. scopulophilus.
Materials and methods
Study area
Leptocereus scopulophilus occupies 3500 m2 of the deep southeastern slope of Pan de Matanzas, a karstic hill located 11 km west the city of Matanzas, Cuba. The population occurs at 212 m above sea level, in the understory of a semi-deciduous tropical forest (Barrios et al., Reference Barrios, González-Torres and García-Beltrán2012). The Pan de Matanzas hill has a seasonal tropical climate with a dry season in the winter. The dry season lasts from November to April with an average rainfall of 331 mm and an average temperature of 23.6°C. The rainy season begins abruptly by May until October with an average rainfall of 1066 mm and an average temperature of 25.8°C. Relative humidity varies from an average of 82.0% during the dry season to 89.4% in the rainy season. Predominant soils are shallow and rocky, belonging to the calcimorphic humic group: red rendzinas, black rendzinas and humic-carbonated (Barrios et al., Reference Barrios, Flores, González-Torres and Palmarola2015).
Species
Leptocereus scopulophilus is a critically endangered columnar tree-like cactus, endemic to the karstic elevations of Mayabeque and Matanzas (Barrios, Reference Barrios2014). The fruits are ellipsoid-globose, 6.5 cm long and 5.5 cm wide, and yellowish-green coloured with persistent areolas at maturity (Areces, Reference Areces1993). Seeds are dark and small, with dimensions reported for the population of Loma de Somorostro of 1.3 mm long and 0.9 mm wide (Areces, Reference Areces2003). Moreover, all seeds are surrounded by mucilage with a pectin matrix (Barrios et al., Reference Barrios, Flores, González-Torres and Palmarola2015). The seeds are dispersed by an endemic rodent, Capromys pilorides (Barrios et al., Reference Barrios, Mancina and González-Torres2012).
Morphological analysis of seeds
In November 2013, ten mature fruits were randomly collected from ten different individuals and 50 seeds of each morph were selected (five seeds of each morph per fruit). The fruit was cut longitudinally spanning only the pericarp (so as to not damage the seeds); both parts of the fruit were rotated in opposite directions until split, and the pulp containing the seeds was removed using a spoon. We considered a mature fruit a light green, turgid fruit with areoles spaced far from one another. Morphs were differentiated using the presence or absence of spermoderma scarification. When describing the seeds, we noted qualitative characteristics such as shape and coat integrity, as well as quantitative characteristics such as length, width, length of the hilum-micropylar region, thickness (Fig. 1) and mass. The measurements were made with a caliper (±0.05 mm) and an analytic Sartorius balance (±0.1 mg).
Traits analysed and regions of Leptocereus scopulophilus seeds. A, side view; B, ventral view. HMR: hilum-micropylar region.
In order to identify morphological differences between the two seed morphs, the seed surface was studied by electron microscopy equipped with Hitachi S-405A and СamScan-S2 after being coated with gold‒palladium in an argonic medium. Photomicrographic analysis used the following sequence of description: (1) spermoderma cell shape, (2) anticlinal cell boundaries (relief, curvature and interstices), and (3) periclinal wall sculpture (curvature and surface). The terminology follows the criteria of Barthlott and Hunt (Reference Barthlott and Hunt2000).
Germination of seed morphs
Two hundred and forty seeds of each morph were collected from eight fruits (30 seeds per fruit for each morph) from eight different individuals. The seeds were selected and washed with running water until the mucilage was removed, as it can delay seed germination of L. scopulophilus (Barrios et al., Reference Barrios, Flores, González-Torres and Palmarola2015). A paired experiment with eight replicates of 30 seeds per morph were sown in Petri dishes (10 cm diameter and 1 cm depth), with two layers of filter paper moistened with distilled water used as a substrate. This experiment was performed in nursery conditions under mesh-filtered light (3515 lx). Day length (irradiation time per day) was 11–12 h and temperature was an average of 24.3°C with a maximum of 32.6°C and a minimum of 20.5°C. These conditions are similar to those recorded during L. scopulophilus germination experiments by Barrios et al. (Reference Barrios, Flores, González-Torres and Palmarola2015). Both light and temperature were recorded every 30 min using a data logger (Hobo) (±0.1 lx and ± 0.1°C) during the experiment. To prevent fungal growth, six drops of 1% sodium hypochlorite were added at the beginning of the experiment according to Naranjo et al. (Reference Naranjo, Rengifo and Soriano2003). Distilled water was added to the saturation point weekly using a sprayer to maintain moisture. For each germinated seed the following features were determined: (1) the seed region from where the embryo emerged (Fig. 1), (2) the primary organ emerged (radicle or cotyledons), (3) the position of the germination cracks, and (4) operculum formation.
The germination behaviour of these two morphs was evaluated by three parameters: (1) minimum germination time (T min), which is the minimum time required for the seeds to start germination; (2) germination rate (T 50), which indicates the time necessary for half of the seeds to germinate relative to the total germinating seeds; and (3) germinability (G) in 28 days, which refers to the percentage of seeds germinated under experimental conditions (Flores and Briones, Reference Flores and Briones2001). Visible emergence of the cotyledons or the radicle through the spermoderma was defined as germination. T 50 was calculated according to Ranal et al. (Reference Ranal, Santana, Resende and Mendes-Rodrigues2009) as:
$${\rm T}_{{\rm 50}} = \displaystyle{{\Sigma n_i \cdot t_i} \over {\Sigma n_i}}$$
where n i is the number of seeds germinated at time i and t i is the incubation period in days.
Data analysis
For all variables analysed, normality of the data was checked using a Kolmogorov‒Smirnov and Levene test for homogeneity of variance. Seed traits T min and T 50 were each compared between seed morphs using a t-test. Germinability and the frequency of germination variants were compared with binary logistic regression (Sileshi, Reference Sileshi2012). All statistical analyses were performed using the STATISTICA version 10.0 program.
Results
Seed morphs
Leptocereus scopulophilus seeds are oval, mussel-shaped and flattened laterally. We define two seed morphs via the integrity of the spermoderma: (1) fragmented seed coats (FSC), which appear scarified (Fig. 2) due to a punctual release of the coat, usually towards the apical tip in both dorsal‒ventral and lateral regions (Fig. 3A), and (2) complete seed coats (CSC) (Figs. 2 and 3B). The seed morphs did not differ significantly in traits (Table 1).
Longitudinal section of Leptocereus scopulophilus fruit showing fragmented seeds coat (FSC) and complete seed coats (CSC).
Leptocereus scopulophilus seeds. A, fragmented seed coat at zone with fragmented spermoderma (ZFE) in the apical end of the dorsal-ventral and lateral regions; B, complete seed coat; C, dorsal-side view of spermoderma surface in fragmented seed coat, anticlinal cell boundaries (Acb), depression (D), isodiametric cells of the lateral region (IC), elongated cells in the dorsal‒ventral region (EC) and striations (S); D, complete seed coat with conspicuous anticlinal cell boundaries (Acb) in the dorsal‒apical region and E: in the border region.
Seed traits of fragmented seeds coat (FSC) and complete seeds coat (CSC) in Leptocereus scopulophilus (n = 50)
HMR: length of the hilum micropylar region.
For both morphs, spermoderma cells are isodiametric and quadrangular to hexagonal on the border and lateral regions. Spermoderma cells in the dorsal‒ventral region, however, are elongated. In this region, the relief of the anticlinal boundaries are channelled and straight, with pronounced craters at the intersections between cells and a convex curvature of the periclinal wall with irregular parallel striations along the cuticle (Fig. 3C). Anticlinal cell boundaries in the border region were variable in both seed morphs. These cell boundaries may either be raised or inconspicuous, in this case with pronounced craters at the intersections between cells. In the lateral region, anticlinal cell boundaries are very hard to define (Fig. 3D and E). Additionally, irregular striations along the cuticle of periclinal walls were more visible in the FSC compared with the CSC (Fig. 4).
Striations (S) of the periclinal wall cuticle of a fragmented seed coat (A) and a complete seed coat (B) in Leptocereus scopulophilus. Dorsal views.
Germination of seed morphs
Four morphological variations were observed where the radicle or cotyledons emerge from the germination cracks in different positions on the spermoderma (Fig. 5). The most frequent variant was operculed hilum-micropylar germination in CSC, which had a different frequency in FSC (Table 2). Operculed hilum-micropylar germination was characterized by the formation of a large dorsal crack (covering 3/4 of the dorsal region of the seed) and two lateral cracks originating near the dorsal edge of the hilum-micropylar region extending to the ventral side of the seed (Fig. 5A). As a result, the hilum-micropylar regions, together with one or more rows of cells in the border region, constitute the operculum, which is pushed during the emergence of the radicle.
Germination variants in Leptocereus scopulophilus. Continuous lines indicate the position of germination cracks occurring in spermoderma. Dashed lines indicate the position of germination cracks on the opposite seed side. A, operculed hilum-micropylar germination; B, dorsal unoperculed germination; C, alternative unoperculed germination; D, alternative operculed germination.
Germination variants of fragmented seeds coat (FSC) and complete seeds coat (CSC) in Leptocereus scopulophilus
n = 8; P-values in bold are statistically significant (at the 5% level)
Dorsal unoperculed germination (Fig. 5B) occurs due to the formation of a large dorsal crack (covering the entire dorsal region of the seed), which is followed by a single lateral crack that originates near the dorsal edge of the hilum-micropylar region and extends to the ventral side of the seed. Thus two valves originate from the seed coats, where the radicle emerges. There is no operculum, as one of the lateral cracks is not formed, and the hilum-micropylar region is attached to one of the valves. In dorsal unoperculed germination, similar to the operculed hilum-micropylar variant, the cracks occur both on the seed coat and on the thin endosperm layer.
Other germination variants were labelled as alternative due to the emergence of the embryo in an area opposite to the point of emergence in previous variants. Alternative unoperculed germination (Fig. 5C) is almost exclusive to FSC (Table 2) and a short crack is formed where the inner integument is exposed at the apical end of the seed. In this germination variant the endosperm and the membrane formed by the inner integument break off and the cotyledons appear through the crack. Finally, in alternative operculed germination (Fig. 5D), a transverse crack is formed at the dorsal region and towards the apical end of the seed, which is continued on both sides until the opposite point of the transverse cracks. As a result, a circumcised apical operculum is formed through which the cotyledons appear.
All germination variants were found in both seed morphs. However, in the CSC, operculed hilum-micropylar germination predominates, while in the FSC, operculed hilum-micropylar germination and dorsal unoperculed germination occurred in similar proportions (Table 2). Moreover, there was a better representation of the alternative germination variants in FSC compared with CSC (Table 2). The minimum germination time (T min) and germination rate (T 50) of the FSC were significantly lower and the germinability (G) significantly higher compared with the CSC (Table 3). Germination was delayed in the CSC and took almost twice as long as in FSC, and the germination rate in FSC was increased by 7 days relative to CSC (Table 3).
Germination of fragmented seeds coat (FSC) and complete seeds coat (CSC) in Leptocereus scopulophilus
n = 8; Test is t-test for T min and T 50, χ2 for G; P-values in bold are statistically significant (at the 5% level). T min, minimum germination time; T 50, germination rate; G, germinability.
Discussion
Characterization of seed morphs
The traits of L. scopulophilus seeds did not significantly vary between morphs. However, it is remarkable that the seeds collected in the population of the Pan de Matanzas hill are double the size in length and width of those reported by Areces (Reference Areces2003) from the Somorrostro Hill population. The results from Areces (Reference Areces2003) may be biased due to the seeds originating from a single individual collected in 1923. However, seed size might be a character of geographic variation in L. scopulophilus, which has been shown to influence the survival of seedlings in several species (Venable, Reference Venable1992; Andersson and Milberg, Reference Andersson and Milberg1998).
On the other hand, the elongated shape of the coat cells of the dorsal‒ventral region agree with Areces (Reference Areces2003), who suggests that these cells in L. scopulophilus are the only exception to the isodiametric pattern that characterizes the spermoderma cells of Leptocereus. Moreover, the visibility of anticlinal walls, and the irregular striations along periclinal walls, provide evidence for the existence of a thin cuticle in L. scopulophilus seeds. This result agrees with Areces (Reference Areces2003), who includes the species in a group with a fine or regular cuticle. We consider that the presence of irregular striations along periclinal walls of the dorsal region in FSC is an indicator of a thinner cuticle in this morph relative to CSC.
Germination of seed morphs
Two seed morphs in L. scopulophilus with different germination behaviour may represent a survival mechanism in the face of environmental variability, as has been suggested by Venable (Reference Venable1985) and Lu et al. (Reference Lu, Tan, Baskin and Baskin2014). This mechanism may act by spreading germination and seedling recruitment over a longer period of time (Lu et al., Reference Lu, Tan, Baskin and Baskin2010; Leverett and Jolls, Reference Leverett and Jolls2014), thus reducing competition between siblings (Cheplick, Reference Cheplick1996) and seedling predation (Janzen, Reference Janzen1970; Schupp, Reference Schupp1988). Currently, seed dispersal of L. scopulophilus covers very short distances, usually under the mother plant, while the germination of its seeds occurs mainly in the rainy season. Given this behaviour, the successful survival of the seedlings of this endemic cactus will depend on the balance between rapid germination (to accumulate sufficient water during the rainy season) and avoiding herbivory, which may be intense at the start of the rainy season and reduced during the dry season (Gonzalez-Torres et al., Reference González-Torres, Barrios and Palmarola2012). Otherwise, having two different germination behaviours may ensure that the population persists in case of seedling establishment failure at a given moment during the rainy season (Venable, Reference Venable1985). In this way, seed heteromorphism in this cactus may act as a response to the environmental variability (temperature, rain patterns, relative humidity and sunlight) to which the population of Pan de Mantanzas is subjected (Gonzalez-Torres et al., Reference González-Torres, Barrios and Palmarola2012).
Four germination variants were observed in the present study for L. scopulophilus seeds, which contrasts with what was reported for the genus by Areces (Reference Areces2003). This author studied the germination of only two species of Leptocereus (L. quadricostatus and L. sylvestris) and their work was fundamentally guided by taxonomic questions, leading them to recognize only the operculed hilum-micropylar germination, which corresponds to the ‘Cereus’ variant described by Bregman and Bouman (Reference Bregman and Bouman1983). A thinner cuticle on the border region could contribute to this variant, since the lateral cracks that release off the operculum (hilum-micropylar region) occur in the basal region according to Bregman and Bouman (Reference Bregman and Bouman1983) and Areces (Reference Areces2003). The ‘Cereus’ variant is the most widespread in the family Cactaceae. According to Vishenskaya (Reference Vishenskaya and Tarhtajan1991), this seed variant is favoured ecologically due to the ability to rapidly germinate in a short period of rain. However, the dorsal unoperculed germination variant represents more than 30% of variants in each seed morph of L. scopulophilus, and it is very similar to the Pereskia variant described by Bregman and Bouman (Reference Bregman and Bouman1983). Therefore the ‘Cereus’ variant might be derived directly from the Pereskia variant and not indirectly as argued by Bregman and Bouman (Reference Bregman and Bouman1983). Moreover, another two germination variants in L. scopulophilus are treated as exceptions by Bregman and Bouman (Reference Bregman and Bouman1983). According to these authors, exceptional variations often occur in seeds of the ‘Cereus’ variant because the lateral cracks that form the hilum-micropylar operculum are not formed. The alternative germination variants occur only if one apical crack is formed prior to the formation of the dorsal crack which initiates the predominant variants. This could be related to the lower mechanical strength at the apical region of the FSC, the seed morph where these alternative variants are well represented. Our results show that the germination variant determines the primary organ of the embryo that is going to emerge. In alternative variants cotyledons emerge first, while in the predominant variants it is the radicle which emerges primarily, as has also been reported by Bregman and Bouman (Reference Bregman and Bouman1983). This is due to the curved shape of the embryo in the Cactaceae family, where the apical end of the seed corresponds to the cotyledons and the border region to the radicle (Bravo-Hollis and Sanchez-Mejorada, Reference Bravo-Hollis and Sánchez-Mejorada1991).
High and fast germination of the FSC compared with CSC provides evidence for physiological differences in germination, as mentioned by Matilla et al. (Reference Matilla, Gallardo and Puga-Hermida2005) for species with seed heteromorphism. Germination differences between the morphs could be related to a thinner cuticle in the FSC and the punctual release of its spermoderma, which facilitates a quick imbibition of the embryo and the coat breaking. The relationship between spermoderma characteristics and germinability for FSC agrees with the idea proposed by Maiti et al. (Reference Maiti, Hernández-Piñero and Valdéz-Marroqumín1994), where a high percentage and quick germination of seeds in Cactaceae may be related to the presence of a thin spermoderma. A similar observation has been reported for Phaseolus vulgaris L. (Powell et al., Reference Powell, Oliveira and Matthews1986), Brassica rapa L. (Puga-Hermida et al., Reference Puga-Hermida, Gallardo, Rodríguez-Gacio and Matilla2003) and Arabidopsis thaliana (L.) Heynh. (Kuang et al., Reference Kuang, Xiao and Musgrave1996). In these species, the seed morphs with thinner spermoderma have a higher rate of imbibition and an increased percentage of germination (Matilla et al., Reference Matilla, Gallardo and Puga-Hermida2005). Moreover, the FSC may originate from changes in development that cause loosening of the spermoderma in certain parts of the coat along with decreased pigmentation. Thus FSC have adaptive advantages in the use of water in drier substrates and may influence the occurrence of vivipary in the seeds of the inner fruit of L. scopulophilus, as reported by Barrios et al. (Reference Barrios, González-Torres and García-Beltrán2012). However, until now, no studies of vivipary in Cactaceae have reported seed heteromorphism or FSC (Conde, Reference Conde1975; Cota-Sánchez, Reference Cota-Sánchez2004; Cota-Sánchez & Abreu, Reference Cota-Sánchez and Abreu2007; Cota-Sanchéz et al., Reference Cota-Sánchez, Reyes-Oliva and Sánchez-Soto2007, Reference Cota-Sánchez, Reyes-Olivas and Abreu2011; Rojas-Aréchiga and Mandujano-Sánchez, Reference Rojas-Aréchiga and Mandujano-Sánchez2009; Aragón-Gastelum et al., Reference Aragón-Gastélum, Reyes-Olivas, Sánchez-Soto, Casillas-Álvarez and Flores2013). Given this evidence, it is not possible to identify the punctual release of the FSC coat as the triggering factor for vivipary in the species. While we have observed that in L. scopulophilus the viviparous seedlings arise from FSC most frequently (probably as a consequence of the majority representation of FSC in alternative variants of germination; see Table 2), we have also found viviparous seeds in CSC, as shown in Fig. 1 of Barrios et al. (Reference Barrios, González-Torres and García-Beltrán2012).
Having a complete spermoderma with a thicker cuticle also has its advantages. This characteristic of CSC can be a barrier to pathogens and increase the resistance to degradation during passage of the seed through the digestive tract of the dispersers. It may also allow for greater seed persistence in the soil, as suggested by Matilla et al. (Reference Matilla, Gallardo and Puga-Hermida2005). While in the present study we did not confirm the viability of the seeds, it is very likely that the 33% of the CSC which did not germinate in 4 weeks are dormant seeds, a trait that is related to the formation of a seed bank (Álvarez-Espino et al., Reference Álvarez-Espino, Godínez-Álvarez and De la Torre-Almaráz2014). Leptocereus scopulophilus possesses permeable seeds that imbibe 25% of their fresh mass in the first 24 h; additionally, tests of viability conducted in our laboratory (both tetrazolium tests and cut tests) demonstrated that the seeds of L. scopulophilus that did not germinate for 4 weeks showed more than 90% viability (authors’ unpublished data). Moreover, Barrios et al. (Reference Barrios, Flores, González-Torres and Palmarola2015), under similar conditions to those in the present study, observed 3‒21% germination of seeds of L. scopulophilus after 4 weeks, which can be considered dormant according to the criteria proposed by Baskin and Baskin (Reference Baskin and Baskin2014). Furthermore, the presence of fruit in the species throughout the entire year (Gonzalez-Torres et al., Reference González-Torres, Barrios and Palmarola2012) may be further evidence that its seeds reach the ground at different times, and at least one transient seed bank is expected. Seal et al. (Reference Seal, Flores, Ceroni, Dávila, León-Lobos, Ortega-Baes, Galíndez, Aparicio-González, Castro, Daws, Eason, Flores Ortiz, del Fueyo, Olwell, Ordoñez, Peñalosa, Quintanar, Ramírez, Rojas-Aréchiga, Rosas, Sandoval, Stuppy, Ulian, Vázquez, Walter, Way and Pritchard2009) published a database which included 83 cactus species with seeds still viable a year after collection and 92 species with seeds tolerant to desiccation. Considering the results of this compilation, as well as the work of Barthlott and Hunt (Reference Barthlott and Hunt2000) and Rojas-Aréchiga et al. (Reference Rojas-Aréchiga, Mandujano and Golubov2013), it can be inferred that the majority of cactaceous seeds that have been studied show appropriate characteristics (small seeds with low mass and positive photoblastism for the persistence of seed banks in soil) (Baskin and Baskin, Reference Baskin and Baskin2014). Although Rojas-Aréchiga and Batis (Reference Rojas-Aréchiga and Batis2001) mention that the formation of seed banks in cacti is unlikely due to the strong predation of seeds, some authors have noted its existence in Echinopsis atacamensis subsp. pasacana (de Viana, Reference de Viana1999), Ferocactus wislizenii (Engelm.) Britton & Rose, Mammillaria grahamii Engelm. (Bowers, Reference Bowers2000, Reference Bowers2005), Opuntia engelmannii Salm-Dyck (Montiel and Montaña, Reference Montiel and Montaña2003), Ferocactus viridescens (Torrey & Gray) Britton & Rose (Angoa-Román et al., Reference Angoa-Román, Bullock and Kawashima2005), Harrisia fragrans Small ex Britton & Rose (Goodman et al., Reference Goodman, Walters, Bradley, Maschinski and Salazar2012), Stenocereus stellatus (Pfeiff.) Riccob. (Álvarez-Espino et al., Reference Álvarez-Espino, Godínez-Álvarez and De la Torre-Almaráz2014), Cereus jamacaru DC. (Santos et al., Reference Santos, Santos, Souza, Andrade, Silva, Andrade and Araújo2015) and in eleven more cacti species of Zapotitlán (Cano-Salgado et al., Reference Cano-Salgado, Zavala-Hurtado, Orozco-Segovia, Valverde-Valdés and Pérez-Rodríguez2012). The possible formation of at least one transient seed bank in the soil by CSC of L. scopulophilus would be an effective strategy for survival during monthly episodes of drought.
Leptocereus scopulophilus is a critically endangered species (González-Torres et al., Reference González Torres, Palmarola, González Oliva, Bécquer, Testé, Castañeira-Colomé, Barrios, Gómez-Hechavarría, García-Beltrán, Granado, Rodríguez-Cala and Regalado2016) with two populations extending no more than a hectare, even though their areas of distribution were probably larger in the past (Barrios, Reference Barrios2014). In accordance with our results we propose the use of FSC for the propagation of the species, given its rapid and high germination capacity. The propagation of the species and its maintenance could be costly, both in terms of introducing it into nature and maintaining it in living collections in botanical gardens. On the other hand, according to Roberts (Reference Roberts1991), establishing a seed collection ex situ would be a more efficient and less costly conservation method, in addition to providing a greater genetic representation of the population when compared with maintaining living collections (Cochrane et al., Reference Cochrane, Crawford and Monks2007). We believe that the use of CSC would be more appropriate in establishing a seed collection ex situ.
In general, seed heteromorphism in L. scopulophilus may ensure successful germination and seedling establishment when considering variable microclimatic conditions and dispersion patterns. In further studies, it would be appropriate to evaluate the seed germination and seedling establishment of both morphs under different conditions (e.g. different water stress and temperatures) as well as the duration of seed viability. This will further clarify the contribution of both morphs to the reproductive success of L. scopulophilus.
Acknowledgements
This study was supported by conservation and research grants from the Mohamed bin Zayed Species Conservation Fund (11052592), British Cactus and Succulent Society. We are particularly grateful to the National Botanic Garden, University of Havana and Planta!-Cuban Society of Botany. We thank Alejandro Palmarola and Luis Roberto González-Torres for the electron microscopy photographs, to Jorge A. Sánchez for useful comments during manuscript review and to Ana C. González for her review of the English version of the manuscript. We also thank four anonymous reviewers and the Associate Editor Costas A. Thanos for their invaluable comments on the manuscript.
