﻿Phylogenetic relationships in Coryphantha and implications on Pelecyphora and Escobaria (Cacteae, Cactoideae, Cactaceae)

﻿Abstract The genus Coryphantha includes plants with globose to cylindrical stems bearing furrowed tubercles, flowers arising at the apex, and seeds with flattened testa cells. Coryphantha is the second richest genus in the tribe Cacteae. Nevertheless, the genus lacks a phylogenetic framework. The limits of Coryphantha with its sister genus Escobaria and the infrageneric classification of Coryphantha have not been evaluated in a phylogenetic study. In this study we analyzed five chloroplast regions (matK, rbcL, psbA-trnH, rpl16, and trnL-F) using Bayesian phylogenetic analysis. We included 44 species of Coryphantha and 43 additional species of the tribe Cacteae. Our results support the monophyly of Coryphantha by excluding C.macromeris. Escobaria + Pelecyphora + C.macromeris are corroborated as the sister group of Coryphantha. Within Coryphantha our phylogenetic analyses recovered two main clades containing seven subclades, and we propose to recognize those as two subgenera and seven sections, respectively. Also, a new delimitation of Pelecyphora including C.macromeris and all species previously included in Escobaria is proposed. To accommodate this new delimitation 25 new combinations are proposed. The seven subclades recovered within Coryphantha are morphologically and geographically congruent, and partially agree with the traditional classification of this genus.


Introduction
Coryphantha (Engelm.) Lem. was described by Engelmann (1856) as a subgenus of Mammillaria Haw. Later, Lemaire (1868) raised it to generic level. Hunt and Benson (1976) proposed Coryphantha sulcata (Engelm.) Britton & Rose as the type of this genus. Coryphantha is morphologically characterized by adult plants with globose to cylindrical stems, covered with numerous spirally-arranged tubercles, flowers that arise at the apex of the stem, stem tubercles with a groove in maturity, and seeds with flattened testa cells (Anderson 2001;Dicht and Lüthy 2005;Hunt et al. 2006). Species of Coryphantha mainly inhabit the Mexican highlands in xerophytic shrublands and grasslands, although some prefer tropical deciduous forests and coniferous forests (Dicht and Lüthy 2005).
The taxonomy of Coryphantha has been complicated. Attributes such as the shape and size of the stem, the number, color, and orientation of the spines change according to the development state of the specimen, causing confusion with members of other genera such as Escobaria Britton & Rose, Mammillaria, and Neolloydia Britton & Rose (Vázquez-Benítez et al. 2016). For instance, Benson (1969Benson ( , 1982 recognized Escobaria as a subgroup of Coryphantha because of the tubercle grooves, an opinion that persists to this day (Zimmerman and Parfitt 2004).
Current infrageneric classifications in Coryphantha have been entirely based on morphology, which has been evaluated according to different criteria, generating artificial classifications. Bravo-Hollis and Sánchez-Mejorada (1991) recognized three series within the genus: Macromeres Britton & Rose, Aulacothelae Lem., and Glanduliferae Salm-Dyck. Lüthy (2001, 2005) recognized two subgenera: Coryphantha and Neocoryphantha Backeb., divided into sections and series. Finally, Hunt et al. (2006) proposed an artificial classification in which three subgenera and three informal groups were recognized. Those proposals have been based on the presence/absence of extrafloral glands at the areole, the type of development and position of the areole on the tubercles, growth form and shape of the tubercle. None of these proposals has been evaluated within a phylogenetic framework.
A previous molecular phylogenetic study of the tribe Cacteae included a few species of the genus Coryphanta (Butterworth et al. 2002). This study suggested that Coryphantha is part of the Mammillaria (=mammilloid) clade, a group that includes other genera such as Escobaria, Neolloydia, Ortegocactus Alexander, and Pelecyphora Ehrenb. The position of Coryphantha within mammilloid clade was further supported by other studies with better sampling and larger molecular data set (Butterworth and Wallace 2004;Crozier 2005;Bárcenas et al. 2011;Hernández-Hernández et al. 2011;Vázquez-Sánchez et al. 2013). Overall, these phylogenetic studies suggest that Coryphantha is not monophyletic (Bárcenas et al. 2011;Vázquez-Sánchez et al. 2013). Recently, Breslin et al. (2021) proposed the recircumscription of the mammilloid clade by recognizing three genera, Mammillaria, Cochemiea (K.Brandegee), and Coryphantha (including Escobaria). However, sampling in the Coryphantha clade was poor. In this study, we performed phylogenetic analyses focusing on the tribe Cacteae to test for the monophyly of Coryphantha and to better understand its relationship to Escobaria. With the phylogenetic hypothesis obtained we evaluated the infrageneric classification proposed by Dicht and Lüthy (2005), and propose the set of morphological characters that define the genus Coryphantha.
Samples of plant tissue from the epidermis and hypodermis of the stem were dried, frozen, and pulverized. Total DNA extraction was achieved by using the DNeasy plant mini kit (Qiagen, California). We amplified chloroplast markers widely used in phylogenetic reconstruction in Cacteae (Vázquez-Sánchez et al. 2013. Specifically, we amplified the chloroplast coding regions matK and rbcL, and the intergenic spacers psbA-trnH and the trnL-trnF (including part of the trnL), and the rpl16 intron. Primers and profiles of thermal cycles followed Vázquez- Sánchez et al. (2013). The PCR products were sequenced at the High Throughput Genomics Unit at the University of Washington (now unavailable). Appendix 1 shows in detail the GenBank accessions for each taxon.
The sequences for each marker were assembled using SEQUENCHER (v. 4.8, Gene Codes Corporation 2007). The matrices were aligned manually with MESQUITE (v. 2.75, Maddison and Maddison 2015). Table 1 shows some numeric records about the taxa and the aligned sequences included in the subsequent analyses. Insertion-deletion events in aligned sequences (indels) were coded using the simple coding method (Simmons and Ochoterena 2000) (Appendix 2). Additionally, eight morphological characters, proposed as diagnostic for Coryphantha and related genera were coded and used in a combined phylogenetic analysis. It has been suggested that in Cactaceae the inclusion of indels and a set of morphological characters in phylogenetics analysis results in more accurate hypotheses Martínez-Quezada et al. 2020). Character states were extracted from the descriptions of the species (Bravo Hollis and Sánchez-Mejorada 1991;Barthlott and Hunt 2000;Dicht and Lüthy 2005;Hunt et al. 2006) and corroborated in the field, in living collections (Jardín Botánico, Instituto de Biología, UNAM), and with herbarium specimens (MEXU). Characters and character states are listed in Table 2. DNA evolution models for each sequence were estimated using the corrected Akaike information criterion (AICc) in JMODELTEST2 (Darriba et al. 2012) on the CIPRES Science Gateway (v. 3.3 Miller et al. 2010) ( Table 1). The Mkv model (Lewis 2001) was assigned for the indels and the morphological partitions. The first matrix was concatenated by including the five DNA regions. The second matrix included the five DNA regions and the indels partition. Finally, the third matrix included the five DNA regions, the indels and morphological characters. A partitioned Bayesian inference (BI) analysis was performed in MRBAYES (v. 3.2.1, Ronquist et al. 2012). The BI analysis for those matrices consisted of two runs of four chains for  TPM1uf+I+G  TPM1uf+I+G  K80+I  TIM1+I+G  TVM+G  -20 million iterations, saving one tree every 1000 generations, and beginning with one random tree. The burn-in parameter was fixed as 25%.
The ancestral states of the eight morphological characters were traced in the selected phylogeny to test them as potential synapormophies of clades. The tracing of characters was performed in MESQUITE (v.2.75, Maddison and Maddison 2015) using the parsimony ancestral state method on the majority consensus tree from the combined BI analysis.
The ancestral state reconstruction (Appendix 3: Fig. A1) showed that the presence of a complete groove on the tubercle (Appendix 3: Fig. A1B), the apical origin of the flowers (Appendix 3: Fig. A1D), the entire margin of the outer tepals (Appendix 3: Fig. A1E), the green color of the fruit (Appendix 3: Fig. A1F), and the flat multicellular sculpture of the lateral side of the seed (Appendix 3: Fig. A1H) were ancestral states to Coryphantha s.s., few or null changes on these characters states occurred inside the clade. In contrast, in the Escobaria clade, the fimbriate margin of the outer tepals (Appendix 3: Fig. A1E), the red color of the mature fruit (Appendix 3: Fig. A1F), and the pitted multicellular sculpture of the seed were ancestral character states (Appendix 3: Fig. A1H). Additionally, growth form was ambiguous in Coryphantha s.s. and Escobaria clade. The absence of glands near the axil of the tubercles was ancestral to Coryphantha s.s., and the presence of those glands evolved independently in two subclades of Coryphantha (Appendix 3: Fig. A1C). In clade II, turgid glands present all year-long were ancestral, while glands present only during flowering season evolved once in subclade D (Appendix 3: Fig. A1C. Finally, watery cortex was ancestral in Corypantha s.s., but it changed into mucilaginous cortex in the subclade F (Appendix 3: Fig. A1G).

Discussion
The close relationships among Cochemiea, Coryphanta, Cumarinia, Escobaria, and Mammillaria have been recognized by several studies (Butterworth and Wallace 2004;Crozier 2005;Vázquez-Sánchez et al. 2013;Breslin et al. 2021). Breslin et al. (2021) recovered them as closely related lineages and redefined their limits. These authors proposed to expand the limits of Cochemiea to include 37 species of Mammillaria, Neolloydia, and Ortegocactus. Our results (Figs 1, 2) recovered, with moderate to low support, the same phylogenetic position of Ortegocactus and Neolloydia. Additionally, Mammillaria mazatlanensis was nested within Cochemiea. Morphological (Hunt 1985) and molecular evidence (Butterworth and Wallace 2004) suggest that M. mazatlanensis is closely related to other taxa now classified within Cochemiea, so it should be transferred (see Taxonomic summary).
In the molecular analysis, Mammillaria sphacelata and M. benecki were recovered, with low support, as the sister group to Coryphantha s.s. In contrast, Breslin et al. (2021) found M. sphacelata to be the sister to Escobaria + Coryphantha. The addition of eight morphological characters in the combined analysis recovered M. sphacelata and M. beneckei within the clade Mammillaria, and supported Coryphantha s.s. and Escobaria as sister lineages. We argue that the low sampling of this early diverged lineage of Mammillaria (Butterworth and Wallace 2004) and the few sequences included do not allow us to conclude about their relationships.
Finally, Breslin et al. (2021) proposed Escobaria and Coryphantha to be a single genus, as traditionally treated by North American botanists (Benson 1982;Zimmerman and Parfitt 2004). However, sampling in Mexican Coryphantha was not representative. Molecular and combined analyses recovered Coryphantha and Escobaria as independent lineages and the ancestral state reconstruction (Appendix 3: Fig. A1) showed that each genus has a unique combination of morphological characters. Our results support the traditional recognition of Coryphantha and Escobaria as separate genera (Taylor 1979;Bravo-Hollis and Sánchez Mejorada 1991;Dicht and Lüthy 2005;Hunt et al. 2006;Korotkova et al. 2021).

Escobaria clade
The eight sampled species of Escobaria, together with Coryphantha macromeris, Pelecyphora aselliformis, and P. strobiliformis form a monophyletic group with high support values (Figs 1, 2). This clade is diagnosed by the tubercles with complete grooves, external tepals with fimbriate margins, and seeds with pitted multicellular sculpture on the lateral side (except in C. macromeris, and Escobaria chihuahuensis) (Appendix 3: Fig. A1, Fig. 3).
Although previous molecular analyses recovered C. macromeris outside the core Coryphantha clade, phylogenetic relationships of C. macromeris were not clear due to lack of resolution (Bárcenas et al. 2011) and insufficient sampling of Coryphantha  (sensu Dicht and Lüthy 2005). Previous morphological analysis of Coryphantha concluded that C. macromeris was the most dissimilar taxon of the genus Coryphantha (Vázquez-Benítez et al. 2016). The main character that differentiates this species from the rest of the species in the Coryphantha clade is the presence of an incomplete groove in the tubercles and fimbriate outer tepals.
Coryphantha macromeris shares the fimbriate outer tepals with the other species of the genus Escobaria (Fig. 3B, C). Interestingly, C. macromeris and Escobaria vivipara show identical flower morphology (Zimmerman and Parfitt 2004). Additionally, E. chihuahuensis shows a shallowly pitted lateral seed coat (Barthlott and Hunt 2000, plate 73.3-4), similar to the flat cells observed in Coryphantha. Probably, the flat sculpture of the lateral side of the seed in C. macromeris is the result of the same development observed in E. chihuahuensis. As observed in Ferocactus (Taylor and Clark 1983) the change of pitted to flat relief of periclinal walls of the seed testa has evolved independently in several lineages of the tribe Cacteae (Appendix 3: Fig. A1H). Given our results, we propose the recognition of C. macromeris as a member within the new rearrangement of Escobaria and Pelecyphora described in the following paragraphs (see Taxonomic summary).
As in previous analysis our phylogenetic hypothesis recovered the two species of Pelecyphora in the Escobaria clade (Butterworth and Wallace 2004;Bárcenas et al. 2011;Vázquez-Sánchez et al. 2013). Traditionally, Pelecyphora is recognized (Boke 1959;Anderson and Boke 1969) by the presence of a rudimentary groove on the tubercles and the "reticulate or striate" seed structure ("par-concave" sensu Barthlott and Hunt 2000). However, Pelecyphora also falls into Taylor's (1979) concept of Escobaria, which is circumscribed by seeds with intracellular pits (par-concave) and grooved tubercles. Following Boke (1959), the rudimentary groove in Pelecyphora (Fig. 3D) is morphologically equivalent to the groove found on the tubercles of Coryphantha and Escobaria. Regarding seed morphology, the pitted appearance of the seed coat in Escobaria happens because only the central portion of the outer wall of the testa cell is thinner and collapses, while in Pelecyphora the entire outer wall of the testa cell is thin and collapses (Barthlott and Hunt 2000). Therefore, Escobaria and Pelecyphora show a pitted lateral seed coat, differing in cell shape and pit diameter.
Finally, the margin of the outer tepals in P. aselliformis may be entire or fimbriate, while in P. strobiliformis is always fimbriate (Anderson and Boke 1969); this character is also observed in all species of Escobaria (Zimmerman and Parfitt 2004;Hunt et al. 2006). We hypothesized that Pelecyphora represents a derived lineage in Escobaria that has changed radically its growth form and the shape of its tubercules to occupy specific niches in the Sierra Madre Oriental. A similar trend is observed in species of the genus Turbinicarpus (Backeb.) Buxb. & Backeb., in which some species have evolved into a globose-depressed growth form with cylindrical and flattened distally (hatchetshaped) tubercles (e.g., Turbinicarpus pseudopectinatus (Backeb.) Glass & R.A.Foster) or pyramidal and dorsiventrally flattened (scale-like) tubercles (e.g., Turbinicarpus schmiedickeanus (Boed.) Buxb. & Backeb.) (Vázquez- Sánchez et al. 2019).
Several studies recovered with high support the alliance of Pelecyphora and a clade including Escobaria tuberculosa, the type species of Escobaria. A diagnostic character of Escobaria and Pelecyphora is the outerperianth-segments with ciliated margins as shown in E. emskoetteriana (Fig. 3B), E. abdita Řepka & Vaško (Řepka and Vaško 2011) and E. sneedi Britton & Rose (Benson 1982) not included in this analysis. The genus Pelecyphora was published in 1843 by Ehrenberg, while Escobaria was published 80 years later, in 1923, by Britton and Rose. In this context, we propose to merge Escobaria members, including C. macromeris into Pelecyphora (see Taxonomic summary) following priority of publication as dictated by the principle III of the International Code of Nomenclature for algae, fungi, and plants (Turland et al. 2018).

Coryphantha clade
Coryphantha can be recognized as a natural group by excluding C. macromeris. Coryphantha s.s. (henceforth Coryphantha) conformed a robust clade (PP = 1, Figs 1, 2) and can be diagnosed by tubercles with a complete groove, flowers with apical origin, outer tepals with entire margin, green fruits, and seed with flat multicellular sculpture on the lateral side (Appendix 3: Fig. A1, Fig. 3).
Although subgenera Neocoryphantha and Coryphantha recognized by Dicht and Lüthy (2005) are partially recovered, our phylogenetic analyses showed that most of the infrageneric sections and series proposed by these authors do not represent natural entities. All sampled members of subgenus Coryphantha were recovered in clade I, including taxa without turgid glands near the axil throughout the year (Appendix 3: Fig. A1C). However, this clade also included two of the species assigned to section Robustispina Dicht & A. Lüthy in the subgenus Neocoryphantha (Table 3), making Coryphantha subgenus Coryphantha (sensu Dicht and Lüthy 2005) a paraphyletic group. Clade II grouped all the members of the subgenus Neocoryphantha section Neocoryphantha, but the members of the sections Lepidocoryphantha and Robustispina ( Fig. 1) were recovered in the clade Escobaria and the clade I, respectively. Therefore, Coryphantha subgenus Neocoryphantha (sensu Dicht and Lüthy 2005) represents a polyphyletic group. All members of clade II show turgid glands at or near the axil throughout the year (Fig. 3K), which is recognized as a consistent diagnostic feature and a potential synapomorphy for this lineage (Appendix 3: Fig. A1C).
In order to reflect the relationships found in our phylogenetic hypothesis and to provide a natural infrageneric classification of the genus, we re-circumscribe the two subgenera in Coryphantha. One for clade I, the subgenus Coryphantha, and another one for clade II, the subgenus Neocoryphantha (see Taxonomic summary). We further propose to recognize the recovered subclades as sections (see Taxonomic summary). The species belonging to each section, their morphological similarities, and their distribution (biogeographic provinces) are discussed below.
Coryphantha subgenus Coryphantha (clade I) emerged in five subclades that partially represent some taxonomic groups proposed by Dicht and Lüthy (2005). However, series and subseries suggested by these authors do not represent monophyletic groups. Clade A included species from series Retusae Dicht & A. Lüthy, Pycnacanthae Dicht & A. Lüthy and Salinenses Dicht & Lüthy (Table 3). In this case, members of clade A present most of the radial spines with subulate shape (Fig. 3F) (Bravo-Hollis and Sánchez-Mejorada 1991;Dicht and Lüthy 2005). Our results found that the species complexes C. elephanthidens and C. pallida do not represent monophyletic groups. This result corroborates that C. bumamma and C. greenwoodii are different species from C. elephantidens as proposed by Vázquez-Benítez et al. (2016). Additionally, our results support the proposal of Arias et al. (2012) to recognize C. calipensis and C. pallida as two distinct species. The distinction of C. pseudoradians Bravo from C. pallida Britton & Rose, remains unresolved, since the former was not included in our analysis.
As documented by Dicht and Lüthy (2005), there was a historical confusion between C. pycnacantha and C. pallida, since they are morphologically similar (Arias et al. 2012). This affinity is now justified since they belong to the same clade. Dicht and Lüthy (2005) classified C. pallida within series Salinensis along with northern species. This species emerged in Clade A, which is recognized here as section Retusae (see Taxonomic summary). This is distributed in central Mexico, encompassing the southern portion of the piedmont of Sierra Madre Occidental, the Mexican High Plateau, the plains and piedmonts of the Mexican Transvolcanic Belt, the southern portion of Sierra Madre Oriental, and the Balsas Basin.
Clade B included members of the series Coryphantha and Corniferae Dicht & A. Lüthy ( Table 3). Members of this clade show upright or radiate tubercles (Fig. 3G). This lineage is recognized in the present work as the section Corniferae. This clade presents a wide distribution and occupies several northern ecoregions. An eastern group of species inhabits the Chihuahuan Desert, the Sierra Madre Oriental, and the Tamaulipas-Texas Semiarid Plain, and a western group occupies the Chihuahuan Desert, the piedmont of the Sierra Madre Occidental, and the Sierra Madre Occidental.
Coryphantha gracilis is classified into the monotypic section Gracilicoryphantha Dicht & Lüthy by the presence of globose seed and broad basal hylum (Dicht and Lüthy 2005). Although C. gracilis was not included in our analysis, we suggest that it belongs to clade B, because of its morphological affinity to C. compacta and C. recurvata (Vázquez-Benítez et al. 2016), and also the similar geographic distribution. Coryphantha pulleineana (Backeb.) Glass was not included in our analysis. Dicht and Lüthy (2005) mention some morphological affinities to C. ramillosa. In addition, C. pulleineana and C. pseudoechinus shared the presence of glands in the spiniferous areole. For now, we propose C. pulleineana as a member of this group because of its morphological and geographical congruence to other species of this clade (Dicht and Lüthy 2005).
Subclade C included two members of the series Salinenses (Table 3). These taxa can be distinguished by the presence of appressed tubercles and woolly stem tips (Fig. 3H) (Bravo-Hollis and Sánchez-Mejorada 1991;Dicht and Lüthy 2005). Our study included C. durangensis subsp. durangensis and C. durangensis subsp. cuencamensis, which formed a monophyletic group. However, they showed different branch lengths, which suggests that its recognition as different species, as proposed by Vázquez-Benítez et al. (2016), must be considered. This small group is recognized in the present work as the section Durangenses (see Taxonomic summary). This group presents a narrow distribution in the state of Durango, inhabiting the Chihuahuan Desert and the piedmont of the Sierra Madre Occidental.
Subclade D corresponds to Coryphantha section Robustispina (Table 3, Taxonomic summary). This clade is supported by the presence of turgid glands near the axil only during the flowering season ( Fig. 3I; Appendix 3: Fig. A1C). Although those species have been grouped in the past with other taxa with glands (Dicht and Lüthy 2005;Vázquez-Benítez et al. 2016), our results suggested that this character state emerged independently from an ancestral with absent glands. This species occurs in the Chihuahuan Desert and in the northern piedmont of the Sierra Madre Occidental.
Subclade E was formed by six taxa classified into the series Coryphanta, series Salinenses, and series Corniferae (Table 3). There are no evident morphological characters that define clade C. Affinities such as the red filaments have been observed in C. echinus, C. kracikii, C. salinensis, and C. sulcata. Particularly, C. salinensis and C. sulcata share a yellow flower with a brilliant red flower throat (Dicht and Lüthy 2005). Also, C. difficilis, C. kracikii, C. salinensis show tubercles appressed, and slightly appressed in C. werdermannii (Fig. 3J). Members of subclade E are proposed here as the Coryphantha section Coryphantha, which is distributed in the Chihuahuan Desert, the Sierra Madre Oriental, and the Tamaulipas-Texas Semiarid Plain.
We propose the division of subgenus Neocoryphantha (clade II) into two sections. The first one is section Clavatae (see Taxonomic summary), which corresponds to subclade F (Table 3). This section presents mucilaginous cortex (Dicht and Lüthy 2005), a character recovered as ancestral to the group in our analyses (Fig. 3K, Appendix 3: Fig. A1G). Section Clavatae occurs mainly in the southern part of the Chihuahuan Desert and in the Mexican High Plateau, with C. ottonis ranging to the interior plains and piedmonts of the Sierra Madre Occidental and the Mexican Transvolcanic Belt. The second is sec- Table 3. Species memberships of the main clades obtained in this study and their previous infrageneric classification by Dicht and Lüthy (2005 tion Echinoideae, which corresponds to subclade G (Fig. 3L, Table 3). This section can be recognized by the presence of watery cortex (Appendix 3: Fig. A1G). Members of the section are distributed in the Chihuahuan Desert and the Sierra Madre Oriental.

Cochemiea
Phylogenetic analyses support the addition of Mammillaria mazatlanensis within Cochemiea. Three lectotypes are proposed.  (Schumann 1901), and the later extension of the description by Gurke (1905) do not indicate that a type specimen has been preserved. Hunt (1985) confirms that a type specimen was not formally designated.  Britton and Rose (1923) chose Engelmann's epithet tuberculosa over strobiliformis, because the last represents a homonym. However, Benson (1969) suggested that the epithet strobiliformis should be preferred over the epithet tuberculosa. Zimmerman and Parfitt (1993+) mention that Escobaria tuberculosa and E. strobiliformis represent two independent entities and the name E. chihuahuensis Britton & Rose should be considered a synonym of E. strobiliformis. Given the difference in opinions, Hunt et al. (2006) explained that the name Escobaria strobiliformis has been incorrectly applied to E. tuberculosa and should be rejected. Hunt (2016) concludes that E. strobiliformis is an inadmissible name or with indeterminate application. In order to maintain the stability of the names listed in this treatment, the name Mammillaria strobiliformis is considered a homonym and should not be applied (Turland et al. 2018 Britton and Rose (1912), the original specimen of Coryphantha cubensis was kept in cultivation at the New York Botanical Garden. A specimen deposited in NY (120678!) whose data on the label coincide with those referred to in the protologue. Elements such as collector and number (J. A. Shafer 2946) and date of collection (1909) coincide with the label of the specimen referred to here, which is why we designate it as lectotype, while the specimen deposited in the US herbarium (1821121 image!) corresponds to the isolectotype.

Coryphantha
Phylogenetic analyses obtained here support the recognition of two subgenera in Coryphantha (clade C1 and clade C2), which are composed by two section (subclade A and subclade B) and five sections (subclades C to G), respectively. Also, 46 species and 12 subspecies of Coryphantha, are recognized. Asterisk (*) indicates species that were not included in the phylogenetic analyses. A taxonomic synthesis is presented. men. Living voucher specimens are identified by their specimen number in cultivation at Desert Botanical Garden (DES), Jardín Botánico, Instituto de Biología, Universidad Nacional Autónoma de México (JB, UNAM), and El Charco del Ingenio, A.C. ND: no data. Table A1. Insertion-deletion events coded in the alignment for each sequence. Deletion=DEL, insertion=INS, simple sequence repetition (SSR).