Ceratozamia is a genus of cycads occurring in eastern Mexico and Central America. In this study, we describe a new species from the Pacific region of Mexico in Guerrero state. This locality represents the most northwestern Mexico distribution for the genus. We focus the comparison of this species with the most geographically proximate and phenotypically relevant lineages for this taxon. We followed an integrative taxonomy approach to evaluate the classification of these species, including geographic location, morphology, DNA barcoding and phenology as primary sources of systematic data. Within the morphological dataset, reproductive structures are described in detail and new characters are proposed for microsporophylls. The comparative morphology of these structures facilitated the elucidation of differences in forms and species for identification. The two chosen DNA barcoding markers – namely, the chloroplast genome coding region matK and the nuclear ribosomal internal transcribed spacer (ITS) region – had low divergence, allowing only 61% of species identification, suggesting slow molecular evolutionary rates. Besides employing these three basic sources of evidence, we introduced phenology as additional information for species circumscription. In addition, this work includes a brief review of the genus at the species-level. This is therefore the most recent review for Ceratozamia across its full geographic range (latitudinal and elevational). Overall, this work further contributes to a comprehensive framework for systematic studies in Mexican cycads.

cloud forest, cycads, Guerrero, integrative taxonomy

The Sierra Madre del Sur (SMS) is a biogeographic province assigned to the Mexican Transition Zone, which includes neartic and neotropical ecosystems (Marshall and Liebherr 2000). This Mexican province runs northwest-southwest parallel to the Pacific Ocean coast throughout 1,200 km. It has great diversity of species, due to its high climatic and topographical heterogeneity (Morrone 2005; Santiago-Alvarado et al. 2016). Its rugged orography has been influenced by ancient geological events. In parallel, regional climatic cycles are relevant to explain its ecological heterogeneity, including overlaps between the Neotropical and Nearctic biotas, both of which house high species richness and endemics (Morrone 2010). The SMS has been divided into three subprovinces, in which the eastern Sierra Madre del Sur subprovince is the most heterogeneous in structural composition. This subprovince includes two districts: the Guerrero and Oaxaca highlands (Morrone 2017).

One of the least represented habitats in the SMS, and perhaps the least studied, is the SMS cloud forest. This type of vegetation is distributed unequally between 600 and 1,800 meters of elevation, and has a very disjointed and fragmented range caused by different climatic cycles (Luna-Vega et al. 1999). Moreover, the cloud forest is the most diverse habitat in Mexico per unit area and is found in small fragmented zones mainly in the Guerrero and Oaxaca highland districts (Rzedowski 1996; Challenger 1998).

The distribution of Ceratozamia Brongn. (Cycadales) is restricted to areas with high humidity in the main mountain systems of Mexico and Central America. The genus occurs in a spectrum of habitats: evergreen tropical forest, oak-pine forest, and cloud forest in the Sierra Madre Oriental (SMO), Sierra Madre de Chiapas (SMCh), SMS proper, and mountains of eastern Central America. The greatest diversity of this genus is found in the SMO. In the SMS, Ceratozamia has only been reported in Oaxaca (Vovides et al. 2004; Contreras-Medina 2016), but the north central portion of the SMS also may have suitable habitats for the genus. This is particularly so in Guerrero, which has small patches of cloud forest and is part of this biogeographic province (Luna-Vega et al. 1999).

Ceratozamia is easily diagnosed by the presence of two horns at the distal end of the sporophylls (Norstog and Nicholls 1997). This genus is one of the most complex morphological assemblages within the family Zamiaceae, because some of its species show considerable intrapopulation variation and most characters commonly used to diagnose species in other cycad genera are polymorphic (Martínez-Domínguez et al. 2017). Taxonomic and molecular systematic studies have contributed to clarify species identification and to provide taxonomic revisions with keys at the regional level (e.g. Pérez-Farrera et al. 2009; Martínez-Domínguez et al. 2016, 2017, 2018a). Generally, quantitative morphological characters provide limited information among closely related species (Martínez-Domínguez et al. 2017). However, reproductive characters have been poorly studied at the species level, both morphologically and in terms of phenological cycles. Detailed studies and descriptions of organs in the reproductive structures – mainly micro- and megasporophylls – could yield potentially useful diagnostic characters in the genus (Martínez-Domínguez et al. 2018a, b).

Recently, during a review of collections of Ceratozamia deposited at the FCME herbarium, our research group found a specimen collected in 1984 from the state of Guerrero. However, this material lacked reproductive structures indispensable for unequivocal identification. Because the previously known biogeographic pattern of this genus was restricted to southeastern and central Mexico, this discovery was a novelty, as the corresponding coordinates would represent the northernmost locality of Ceratozamia for the Mexican Pacific. Given that similar specimens from the same geographic point had not been collected again, we explored the corresponding area in search of this underdescribed population. After conducting extensive fieldwork in Guerrero to collect fresh material and monitor the attendant reproductive process, we compared the new specimens with all known species in the genus with an initial focus on similar species using geographic and morphological criteria. Formally, we adopted an integrative taxonomic approach (sensu De Salle et al. 2005; Goldstein and DeSalle 2010) that involves a set of inferential rules for corroboration or refutation of species hypotheses based on multiple sources of evidence (‘taxonomic circle’ sensu DeSalle et al. 2005). Elsewhere, we have successfully used this approach for species-level identification in Ceratozamia, explicitly considering intraspecific morphological variation (Martínez-Domínguez et al. 2017). To identify this unidentified herbarium sheet and additional specimens from two Guerrero populations, which ultimately turned out to match the new taxon herein described, we use geographic location information, both qualitative and quantitative morphological characters, character-based DNA barcoding and phenological data. Finally, we have explored how this taxonomic discovery might alter our understanding of biogeographic and evolutionary patterns in Ceratozamia.

In this synthesis of Ceratozamia, we constructed a database for the genus based with geographic, morphological, and molecular data through the use of comparative and character-based DNA barcoding methods. In addition, we investigated the potential value of phenological reproductive patterns for species delimitation; this type of ecological information has been scarcely studied in cycads (Stevenson 1981; Clugston et al. 2016) or in integrative taxonomy studies of gymnosperms. Considering the different male plants in a population, the phenology of pollen strobili shows a broader cycle at the population level; the lifespan of a pollen strobilus is shorter than for ovulate strobili, and an open pollen phase is prolonged in different individual plants (Martínez-Domínguez et al. 2018b). In contrast, duration of ovulate strobili exhibits high specificity in the receptivity phase, which suggests that these data are taxonomically informative. Our findings on the length of the lifespan of reproductive structures and phenological phases coincide with our previous observations in C. tenuis (Martínez-Domínguez et al. 2018b). Ceratozamia leptoceras has a different phenological pattern for ovulate strobili compared to its morphologically similar congeners, which perhaps represents a reproductive barrier in the field (Fig. 5). Unfortunately, these data are not available for all Ceratozamia species.

Despite the species diversity of Ceratozamia and its restricted geographic distribution, a formal infrageneric classification for the genus has not been proposed. This is mainly because previously reconstructed phylogenetic relationships display remarkable differences in the number of clades and the members of the clades (cf. González and Vovides 2002, 2012; Medina-Villarreal et al. 2019). This situation has led to an incorrect identification of specimens that unfortunately has limited the inference of evolutionary relationships, with contradictory results inexplicably unrecognized (cf. Medina-Villarreal and González-Astorga 2016; Medina-Villarreal et al. 2019). The groups within the genus are based on similarities in geographic distribution and morphology (Vovides et al. 2004); however, Ceratozamia species complexes defined by these authors are more influenced by geography. In this work, we follow the proposal of Stevenson et al. (1986) to address morphology in this genus. Our morphological description for all Ceratozamia species recognized tree groups based on the affinity of morphological characters, both vegetative and reproductive, which facilitates comparisons between species.

Under integrative taxonomy criteria (sensu De Salle et al. 2005), we have proposed a new Ceratozamia species from the Sierra Madre del Sur. This biogeographic province is a salient area in terms of gymnosperm diversity in the country (Contreras-Medina 2016). This new record for the genus in Guerrero represents an expansion of the distribution pattern for the genus on the Pacific seaboard, which opens new questions on the influence of mountain systems on the diversity of the Mexican flora. Our survey shows that the vast majority of Ceratozamia species grow at relatively high elevations (Fig. 2); the genus has standard elevation gradient patterns and does not exhibit occurrences at higher elevations or lower elevations. Therefore, it appears that Ceratozamia is poorly adapted at lower and very high elevations (Fig. 2). Species with the lowest elevation distributions occur in southern Mexico and Central America, and those at higher elevations occur in the SMO, Highlands of Chiapas, and SMS. This atypical distribution pattern seems to be favored by mountain regions, mainly in areas with cloud forest.

Historically, taxonomic research in the genus Ceratozamia has been characterized by difficulties in species identification (Miquel 1868). Recently, new species have been described (e.g. Martínez-Domínguez et al. 2018a) but only some of them have been evaluated. Particularly, in relation to DNA datasets, only 19 species could be diagnosed with ITS+matK. Other loci already tested (psbK/I, rpoC1, rbcL and atpF/H) have not provided enough variation for improved resolution (see Nicolalde-Morejón et al. 2011; Martínez-Domínguez et al. 2017). As an approach to molecular species identification, character-based DNA barcoding is susceptible to the addition of new sequences (Nicolalde-Morejón et al. 2010); therefore, further sequencing of new loci (e.g. single-copy nuclear genes; Salas-Leiva et al. 2014) could contribute new molecular diagnostic characters and improved resources for automated species identification in Ceratozamia and other cycad genera.

In terms of morphological taxonomic evidence, vegetative characters have been widely used in the genus to identify species (Vovides et al. 2004; Whitelock 2004; Pérez-Farrera et al. 2009). However, these characters are polymorphic in some species, which hinders the construction of dichotomous keys (Martínez-Domínguez et al. 2017). The classification of strictly allopatric taxa, as is the case in some Ceratozamia species, will always remain arbitrary to some degree so that the evaluation of new characters is a pressing task. In this work, we introduced previously unreported reproductive characters such as the internal parts of the strobilus, which have not been described before in detail for the genus Ceratozamia (Fig. 4; Table 1). Comparisons of pollen strobili and the microsporophylls among different species indicated strong morphological divergence limited to certain groups of related species (Fig. 4). For this reason, we only present a key for the species group that includes the newly described taxon. Thus, it is necessary to explore these reproductive characters for the rest of the species in the genus. Keys based on vegetative features still preserve their relevance due to the dioic condition of these gymnosperms, and the lack of reproductive structures in materials deposited at the consulted herbaria (Morettii et al. 1989). However, we consider that, in view of the complexity of vegetative morphological variation, keys based on reproductive characters could function as excellent support tools in the systematics of cycad genera such as Ceratozamia.

Species delimitation in Ceratozamia exhibits a high degree of complexity (Whitelock 2004). The morphological and molecular evidence available for the genus at present does not allow the diagnosability of all species. For example, C. hildae is a morphologically distinctive species, but lacks molecular diagnostics; in contrast, C. vovidesii and C. mirandae are species that share many morphological affinities, but variation at the molecular level facilitates their taxonomic differentiation (Table 3). This taxonomic-systematic scenario points to the need to explore new molecular markers and the evaluation of variation between populations with morphological, molecular, and phenological evidence for some closely related species. In particular, there are some populations in Oaxaca and Chiapas that still await closer inspection. In the present work, 31 species have been recognized for the genus (Stevenson et al. 1986; Nicolalde-Morejón et al. 2014; Vovides et al. 2016; Pérez-Farrera et al. 2017; Martínez-Domínguez et al. 2017, 2018a), with C. becerrae and C. microstrobila treated as synonyms of C. zoquorum and C. latifolia, respectively (Stevenson et al. 1986; Martínez-Domínguez et al. 2017, 2018a).

This study adds to recent research that suggests the significant role of topography in SMS as a speciation driver in shaping its high species diversity (Santiago-Alvarado et al. 2016). The geographical barriers and ecological changes in this biogeographic province could have allowed the intermingling of different species; these processes could have favored the development of a number of centers of endemism (Morrone 2010; Santiago-Alvarado et al. 2016).

Our integrative taxonomic assessment provided support for the recognition of a new species, Ceratozamia leptoceras. However, the taxonomic complexity of the genus indicates the need of further systematic revisions using multiple sources of evidence, particularly in some groups of species with problematic boundaries. In addition, we have demonstrated the value of investigating ovulate and pollen strobili – particularly, microsporophylls – for the construction of refined morphological matrices for Ceratozamia; and finally, that the construction of dichotomous keys with vegetative characters should consider variation at the population level.

The first author thanks Posgrado en Ciencias Biológicas (UNAM) and CONACYT for the award of a Doctoral Thesis scholarship (729810). The authors would like to express their deepest gratitude to “Mixtecos” people of San Pedro Cuitlapan for providing all the support throughout fieldwork. The authors acknowledge Jesús Ricardo De Santiago Gómez, Luis Tlaxcalteco Tepo and Guadalupe Hernández Martínez for their assistance during fieldwork. We are also grateful to the curators of herbaria consulted for access to study of their specimens (CIB, FCME, MEXU, NY, and XAL), and especially to the curators and staff of ENCB and US of Smithsonian Institution for loans that allowed direct specimen study. Laura Márquez and Nelly López (LANABIO, Instituto de Biología, UNAM) kindly performed DNA sequencing. Finally, we thank Dr. David Gernandt (IBUNAM) for his valuable comments and editorial suggestions on early versions of this paper. This work was financed by Instituto de Investigaciones Biológicas (Universidad Veracruzana, Mexico) and by NSF Grants BSR-8607049 and EF-0629817, awarded to DWS, and logistically supported by FVS.